Yu-Chong (YC) Tai, Electric Engineer and Medical Device Innovator

By David Zierler, Director of the Caltech Heritage Project

March 8, 16, 20, 2023

When the first exploratory conversations began about creating a new department of medical engineering at Caltech, one unavoidable concern centered around competitiveness. At peer institutions, medical or biomedical engineering had been well established, with programs going back decades. But YC Tai saw opportunity in transforming a Caltech truism into opportunity. The frontier is where Caltech shines, and he was convinced that its entrance into the field as a newcomer was an asset, not a liability. Over ten years later, and Tai's confidence has proved prescient: the Department of Medical Engineering and the MEMS Lab have pioneered a vast array of medical research, diagnostics, and devices, and it is a powerhouse of collaboration that connects Caltech to government and academic institutions and a wide array of startup ventures that are translating ideas in the lab to clinical and therapeutic products.

In the discussions below, Tai reflects on his family background and upbringing in Taiwan, and the opportunities that brought him to UC Berkeley for graduate school, where he focused on micromachines and VLSI. In a narration that matches the origin story of so many Caltech faculty appointments, Tai explains why he was drawn to join Caltech, where he was so impressed with its record of support for junior faculty, its research culture that minimized distinctions between disciplines, and a mandate only to do great work with the assurance that everything else will fall into place.

As Tai explains, the launch of the Medical Engineering department was as much a formalization as a creation, spurred in part by past president Jean-Lou Chameau, who emphasized Caltech's capacity for societal impact. Tai surveyed the landscape around the EAS Division and recognized that numerous research groups were engaged in medical and translational research already. The creation of the department would serve as a "force multiplier" to make the sum of this disparate work even more interdisciplinary and impactful. Looking to the future, and after recently stepping down from positions of administrative leadership, Tai conveys optimism for breakthroughs and discovery that will continue to define the field so that the ailments of today become the therapies for recovery and health tomorrow.

DAVID ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Wednesday, March 8, 2023. I am delighted to be with Professor Yu-Chong Tai. YC, it's great to be with you. Thank you so much for joining.

YC TAI: Thank you, David. It's my pleasure to talk to you about me. This is very different than any other conversation I've had.

ZIERLER: I hope it will be fun for everyone involved.

TAI: I hope so, too.

ZIERLER: To start, would you please tell me your title and institutional affiliation here at Caltech?

TAI: My title is the Anna Rosen Professor of Electrical Engineering and Medical Engineering. The division counts me half and half, 50% in electrical engineering, 50% in medical engineering.

ZIERLER: At Caltech, in the division of EAS, these are departments as they exist in a traditional academic setting. This is unique. At other universities, they have departments. Outside of EAS, we have options. But here, it's actually a Department of Electrical Engineering and a Department of Medical Engineering.

TAI: Correct. It happened roughly around 10 years ago. I don't remember exactly which year. Actually, the reason is very simple. EAS is bigger and more complicated than other divisions. There were 14 options. It was too hard to manage. There was a division of the EAS office to establish this department structure to reduce that to seven, which would make it much easier to manage. They also believe this way is more efficient in terms of teaching and hiring arrangements. We ended up with seven departments.

The Creation of Medical Engineering at Caltech

ZIERLER: Your dual appointment speaks to the fact that you're both, of course, an electrical engineer, but your focus is in medical engineering.

TAI: Correct. Actually, I was with electrical engineering always until 2012. The decision was made–actually, there was a suggestion, including by me, I was one of the original 10 faculty to suggest that with the EAS division, we should establish the option of medical engineering. And in 2012, I received a phone call from Professor Ares Rosakis from ME and Aerospace. He fully supported me starting this new option. And then, at that time, because we already had a department arrangement, there was no other options that should've merged with medical engineering, so medical engineering itself became a department.

ZIERLER: Were the Cherngs already involved? Is that what sort of precipitated this?

TAI: Actually, no, the Cherngs got involved in 2016. And then, I think the formal endowment was established starting in 2017. In 2012, the decision was already decided to start a department with absolutely nothing. Zero dollars, no resources, no secretary. It was purely strong intention and the justification that Caltech should have a department of medical engineering. I actually did a presentation to the board of faculty in 2012, and at the end, there was a vote from the board of faculty as to whether to start this new department.

ZIERLER: You were obviously successful. What did you advocate? What was your selling point?

TAI: It was successful. In fact, the final vote was unanimously yes. There were zero no votes. I told the faculty, "Look, biomedical engineering is a continuously growing field. If you take a look at all engineering, including mechanical, materials science, what have you, in the whole world, there's one department that's continued to grow over the last 80 years, and it's biomedical engineering. All other engineering disciplines peak and start to drop. Some departments stay very low and never become glorious. But biomedical engineering–this is from National Science Foundation numbers–continues to grow. The desire to develop new medical technology and to treat diseases continues to grow, even among young people."

ZIERLER: What do you think accounts for the trend? Why is biomedical engineering so unique in that regard?

TAI: I think a lot of young people are influenced by disease in their friends and families. A lot of young people actually also believe modern technology gives them a lot of opportunities to improve technologies and improve medicine further. There's another big reason, too. If you take a look at industry sectors, you will find out medicine, which includes drugs and medical devices, continues to grow. People usually pay attention to Silicon Valley. It's amazing. If you check companies in Silicon Valley 60 years ago compared to now, there are only a few companies left. In fact, In most industries , companies come and go. But you don't see that in medicine. In medicine, once you make it–it's difficult because of the FDA's bar is very high so once you make it, the company seldom disappears. Think about all the drug companies, or medical equipment companies, or medical device companies. They never disappear. They continue to grow, and with a huge growth rate. And profit margins are always high. Because when you're poor, you still need to treat your disease. When you're rich, you want to treat your disease even more. It doesn't matter whether the economy is good or bad. Medicine will continue to grow. And in my class, I show that if you plot, since after World War I, the percentage of spending for each household, there are certain major categories. Housing is one. [Laugh] and there are food, transportation, etc.

ZIERLER: The really important stuff.

TAI: But the percentages people spend on housing, entertainment, transportation, food, etc. are more or less the same. There's, however, one category that continues to grow unbelievably, from zero to 17 or 18% today, and that's medical spending. The United States today spends at least 16% of GDP on medicine and healthcare. There's no other single category that's even close. Even DOD spending is much less than that. That's how advanced countries, when you look at the United States and Europe, care about health. All these reasons add up. You talk about young people, and they want to be doctors, lawyers, businessmen, so as to make a lot of money. But doctors are always there. This industry continues to grow. And medical companies never disappear. Maybe there are mergers/acquisitions, but they never close their doors or go bankrupt, unlike any other sector. That's a fact.

ZIERLER: You use the term biomedical engineering, and the department is medical engineering. Is there any distinction, or do you use them interchangeably?

TAI: Actually, when we started this department, we looked around. We looked at Harvard, Johns Hopkins, and other schools to see what kind of programs they had. A lot of schools use the term bioengineering or department of bioengineering, but they don't have biomedical engineering, and vice versa. But it's interesting that Harvard has both. Harvard has a department of bioengineering and a separate department of biomedical engineering. The reason is very simple. If you look at all these schools, in a way, under the big umbrella of BME, which we assume includes bioengineering and medical engineering, you'll find there are two directions people go. One direction, they actually do bioengineering to study biology. It's very scientific and very close to biology, but they emphasize engineering approaches, applied engineering technologies, to study biology, not medicine.

On the other hand, there are departments, like Case Western Reserve's BME program, which is geared towards taking care of diseases. They don't emphasize biology as much. So there are two directions to go. One is to study the biology of health science. The other is to solve disease problems, to treat people, and to make patients feel better. Then, if you take a look at Caltech, bioengineering was started within EAS 10 years before Medical Engineering. And I was one of the faculty in the bioengineering program. But it's geared towards biology. We have a need or opportunity that there's a lot of interest from faculty, to create another group to solve medical problems, to treat diseases, get close to patients, and to work with doctors.

ZIERLER: Were you doing all of this before the creation of the department? Was your research agenda focused on medical engineering, and you needed a department to surround yourself?

TAI: Yes. Actually, I show slides to the faculty board when I do the presentation. It was not just me. Many faculty already feel that, "We're already doing research naturally." Why? Because the federal government put a lot of research money into biomedicine. Naturally, a lot of our research can be applied to biomedicine. Even before we started the option or the department, many faculty were already doing research around biomedicine.

ZIERLER: There was almost a department here, except it wasn't formalized.

TAI: That's right, it wasn't organized. It was individual effort. Some faculty would suddenly publish a paper, for example, to find out some biomolecule that can be a drug candidate. Or an engineering faculty would find an OCT that can image cells but sporadically and unorganized. A few faculty gathered together, made a few phone calls, had a lunch at the Athenaeum, and we said, "We've got to propose this. We should be organized." We then wrote a white paper and said, "We should formally do this. We should get organized." Even 10 years before 2012, very shortly after I joined Caltech as an electrical engineering faculty, I was in electrical engineering but my research naturally moved towards biomedicine because I made small things, very small electrical, mechanical, and chemical things.

Obviously, they have a lot of biomedical applications. I got funding from NIH, which funds healthcare things. I got funding from DARPA, which cares about soldier health. Naturally, we'd go into biomedicine. Finally, in 2012, I received a formal, "Go do it," from the division. We made a presentation to the board saying, "Biomedical engineering is something we cannot ignore. Students are interested, the federal research money is rich. And also, all our peers, MIT, Berkeley, Stanford, all had biomedical engineering probably 100 years earlier than Caltech. [Laugh] We've got to do it. For the students, for faculty, and for research." Another good timing was that President Chameau took over after President Baltimore. One thing on his agenda he started to talk to faculty about was Caltech's impact to society.

ZIERLER: More of an emphasis on applications.

TAI: Societal impact. And then, I started to say, "Look, if we start a new department of medical engineering, develop technology, directly treat patients, improve healthcare, that's quite a social impact opportunity we have."

ZIERLER: Do you think even before Jean-Lou Chameau, David Baltimore, as a biologist, was the way he elevated biology relevant?

TAI: It's relevant, but President Chameau made it 100 times more important. The timing was right, a lot of faculty were already doing it, and it didn't cost Caltech anything to organize it. "Why not do it?" There were a lot of questions asked. For example, "If you start it now, you're 120 years behind Harvard. [Laugh] How do you compete with them? Johns Hopkins and Harvard both have medical schools. How will you compete?" I told them, "Look, hear me out. You think we're 120 years behind because we're 120 years behind on the old medical technology they've been doing forever. We don't do that." I told my fellow faculty, "If you take a look at Caltech's faculty, every one of us works on either micro- or nanotechnology." "We all have wonderful technology, but in the past, we've only used it for electrical engineering, mechanical engineering, chemical engineering, etc. We never used those wonderful technologies for medical engineering. Now is a turning point. If we start applying our knowledge and our first-rate technology at Caltech, towards medicine."

I told my friends and colleagues, "Harvard and Johns Hopkins need to be scared. They need to catch up with us. [Laugh] We don't need to catch up with them. Don't mix it up. We don't do their old things, we use new technology for new applications. We're going to lead." Then, they said, "We don't have a medical school." I said, "That's an advantage." Why? If Caltech had a medical school, it would be very small. Then, naturally, we'd be trapped, tied up with just a few people at Caltech. In fact, we observe that at most other universities. Engineering faculty usually only work with their medical school people. I said, "Without a medical school, we're going to work with everybody, the best people in the nation." They said, "Why would they work with you?" All these questions. I said, "Listen, do you believe we're the first-rate engineers?" "Of course, we're first-rate engineers at Caltech." "Don't you think doctors want to work with first-rate engineers?" "Absolutely." I told my fellow colleagues, "They're going to come knocking on our doors. Trust me on that." If I'm a doctor, there's a problem, I have an idea to solve it, and I need engineering, who do you go to?

ZIERLER: The best engineers.

TAI: You got it. Even after just a few years–we formally started the department in 2013–people were knocking on our doors. There's another thing I did. I started calling friends at City of Hope, UC San Francisco, UCLA, USC, and said, "Hey, we have a new department. We have smart faculty, smart post-docs, smart graduate students. There's a chance we can work together. Do you want to have a workshop where we can brainstorm together?" We started to do this kind of meeting. Medical people present their problems and their ideas on how to solve the problems, but they need engineering. Then, the engineers sit there and say, "Maybe I can do this and that with you." Very quickly, we found our capacity all taken up. [Laugh] We couldn't take more because we're so small and limited now. Every faculty in medical engineering pretty much says the same thing, "too much demand, not enough capacity to suffice everybody." That's our situation. It's a really good problem to have. But it's exactly what I predicted when we started the department.

ZIERLER: Do you see yourself purely as an engineer? Is there anything you do that is science-oriented?

TAI: I actually call myself a medical device person. In biomedicine, there are two categories. One is drug, such as in BBE. People are actually very deep into molecules, biomolecules, and all that, they naturally go towards drugs. But engineers naturally go toward many medical devices. In fact, the size of both industries is bigger than semiconductors. Most people don't understand that. Also, much bigger than automobiles, much bigger than aerospace. These are real numbers. Medical devices alone are $500 billion a year. Drugs is even bigger, more than $1 trillion a year. It's a big industry. It's very healthy, and companies don't go away. Even with a bad economy or stock market crisis, biomedical companies stand. You'd never hear, "Merck is closing their doors." Are you kidding me? No. [Laugh] Medtronic? No way. And government spending on healthcare continues to go up. It's the single biggest category. All this supported Caltech to start medical engineering. We already had bioengineering and those people do biology, fine. We do diseases and medical devices.

ZIERLER: Within medical engineering, is there anything you don't do? Do you have a research focus? There are so many things that you're a part of. Is there anything that you exclude?

TAI: No, we never set limitations because this is Caltech. We believe smart brains should continue to think of breakthroughs, think out of the box, break boundaries. Why would we set any limitations? At Caltech, there's academic freedom wherever allows it, so we never set any limitations. We just tell everyone, "Do something new. Do something great." If someone says, "Good, it's great," do it. No limitation.

ZIERLER: What are you working on right now?

TAI: I'll give you some examples, some crazy ideas. That's what Caltech people do, including myself. The student I'm going to graduate soon, for example, we're trying to cure type-1 diabetes. Type-1 diabetes, the reason it's a problem is because the beta cells, in the pancreas and the islets actually die. It's a self-immune problem where their lymphocytes attack their own pancreas. The idea is grafting, meaning harvesting the pancreas tissues, or even using stem-cell-derived beta cells. When glucose comes in, those cells produce insulin. If we implant those cells into someone with type-1 diabetes, it can cure type-1 diabetes. I'm not the first to do this. People have tried this for decades, and no one has found a solution yet. When you implant foreign cells into a body, the immune system is going to attack and kill them. We're working on a way to do two things. One, to keep this implant happy and healthy in terms of having enough oxygen and nutrients. We actually published a recent paper where we were the first ones to guarantee there's no lack of oxygen, no ischemia. The second thing is, we're designing devices to prevent your body's immune cells from attacking those cells.

ZIERLER: You're tricking them.

TAI: We literally make cages so the immune cells cannot go in. But oxygen, nutrients, insulin, glucose can go in and out freely.

ZIERLER: What does the cage look like? How does it do it?

TAI: We call it a mesh bag. It's really a bag with a lot of small holes. But the bag wall, made of mesh, is empty. The mesh is hollow, so oxygen can flow freely to guarantee that every cell gets a fair share of oxygen. We've done animal experiments on rats and mice, and it shows really encouraging results.

The Field of Microfluidics

ZIERLER: Tell me about the field of microfluidics. How far back does it go, how early did you get involved?

TAI: As early as 30 years ago, actually. When I joined Caltech, I immediately also joined mechanical engineering and collaborated with someone in fluid mechanics. Because I realized already that making small devices can have the biggest impact in microfluidics. In fact, when DARPA started the microfluidics program, I was one of very few guys in the nation getting funding from DARPA to work on microfluidics. We published a lot of papers in the domain of microfluidics, including making microvalves, how to control them, how to do mixing. For example, I'm one of early researchers in the nation, who has published microfluidics-based high-performance liquid chromatography from a chip. I hope someday it will become a real product. We were also funded by NASA for eight years to develop one technology they wanted to put in the International Space Station.

ZIERLER: What was the technology?

TAI: It's very interesting, it was a complete blood count technology. Right now, there's very minimal medical technology in the International Space Station. On Earth, they draw your blood, they assay, they test. If you want to do a complete blood count, there's a complicated assay. But that's with most basic technology on Earth, big machines running. They can't put that in the International Space Station. They want everything portable, very small, runs on its own. That's microfluidics. We developed a technology where we only need one drop of blood to do a complete blood count. You put it in something like a glucose test paper, small device. You put one drop of blood in there, and it'll tell you your complete blood count. It's targeted for FDA approval this year. That was one graduate student who did his PhD and started a company, tried to "maturize" that technology, and target it to be used on Earth. And you never know, maybe it can be qualified for space and someday for astronauts. Elon Musk is talking about sending people to Mars. What kind of medical technology will they need? That's a big question. They should build a hospital there before they send people there. Otherwise, it's murder, right? I argue that's the case.

ZIERLER: Tell me about lab-on-a-chip. What does that mean?

TAI: Well, this technology to do a complete blood count with one drop of blood is a perfect example of lab-on-a-chip. In fact, before lab-on-a-chip, the even bigger goal before that was hospital-on-a-chip. Then, you understand that lab-on-a-chip is a very small part of hospital-on-a-chip. Lab-on-a-chip means you analyze maybe one drop of fluid, what's in there, in terms of molecules or pH value. Nothing directly related to medicine yet. Whatever people do in the lab. You measure conductivity of fluid, measure biomolecules a cell produces. Today, one cell comes in, you analyze the DNA sequence of the cells, or messenger RNA, or what kinds of proteins and how much are in the cell. Those are all lab-on-a-chip. But hospital-on-a-chip is different, but bigger and grander. That will actually detect and treat your health condition. For example, maybe there's a genetic disease that can be analyzed.

I have another project, using microfluidics to grow cells, and the cells can produce antibodies, and that has medical applications. You don't want a whole hospital, you just want a hospital on a chip. You do something, then throw it away, such as to synthesize a drug. Today, we take an Aspirin pill from a bottle. In the future, it can be a microfluidic chip on demand. You say, "I want Advil," and it will synthesize Advil right there for you. Those are noble goals for the future that microfluidics could do. From the science, there's no reason it cannot be done. Science doesn't prevent this. There are other things science will tell you, "No, it doesn't work." [Laugh] But here, science says yes. A lot of research can be done. But it takes time, money, and crazy scientists to really get it going.

ZIERLER: Are you focused on bringing devices to market, or do you develop the technology and license this out to biomedical companies?

TAI: I really enjoy teaching and research. I like to work with smart post-doc and graduate students, but I set a goal for every project that could lead to real medical devices that someday could be used in every household or hospital. But I draw a line that a student should be qualified to get a PhD even if they just prove the feasibility or solve the most difficult part. They've already created a new technology. They should always graduate in five years. They should never spend their lives here at Caltech. I draw that line very clearly. But then, usually, the research leads to patents, which always go through Caltech. Caltech owns all the patents. And if the students or anybody else is interested in that technology, they can license it from Caltech and start a company. And I may be involved as a consultant. I'll never quit my job here to run a company. That's not my interest. But it turns out, if you work on some new technology and target for new medical problems that haven't been solved before, people will always be interested in commercializing it.

ZIERLER: Because of your interests in academia, are your graduate students and post-docs more likely to go into faculty positions than might be the case at Berkeley or Stanford, for example?

TAI: Both. I've been with Caltech since '89, 34 years. I've graduated 50-some students, and more than 10 are faculty. It's their personal choice. They enjoy research and all that. And they also do the same thing, produce interesting patents and technologies and spin off companies. I have probably more students interested in commercializing whatever PhD they do. They become CEOs and CTOs of companies. I never interfere. But for myself, I draw a line. I stay at Caltech. I enjoy doing research and working with smart, young people.

Startups and Perspective on the Regulatory Framework

ZIERLER: You're not really involved with the FDA and regulatory processes. That's sort of out of bounds for you.

TAI: The companies that spin off from my lab deal with those issues. I don't. I don't want to waste my time with it because I'm not good at it. [Laugh] I only want to spend time doing what I'm good at. And I've been telling people, "Don't ask me to be the CEO of a company. It would be a disaster because I'm not good at it." I think I'm good at thinking of great ideas, working with students, doing research, publishing papers. I enjoy that, and I think I'm good at it. I'm not good at other things. [Laugh]

ZIERLER: What would be an example of an idea that you develop in the lab that a student then takes to develop a company?

TAI: One drop of blood to get a complete blood count is a perfect example.

ZIERLER: What's the company or the outcome of that research?

TAI: Well, that company's called CytoChip, and they really perfected the tiny machine, and then they designed a microfluidic cartridge such that if you put one drop of blood in there, it will tell you the complete blood count. That means how many white blood cells you have, how many kinds of white blood cells you have, all that. That's by definition a complete blood count. They're dealing with the FDA, not me. [Laugh]

ZIERLER: Just the way you like it.

TAI: Yeah, just the way I want it. Hopefully, they can get FDA approval at the end of this year. That's one example. I also have other examples. There's another company where the graduate student learned the silicone material technologies, developing contact lens drug delivery. There are a lot of eye diseases that can be dealt with that way. There's a big difference, topical drug delivery usually is a one-time kind of thing. You put the liquid drop on the eye, the drug is there, and very quickly it disappears. But research shows there are some eye diseases that small amount of drugs should be applied continuously, and the contact-lens drug delivery addresses exactly that.

ZIERLER: Constant contact.

TAI: Yes. Much less drug, but continuous delivery. It treats the disease better. All that actually could happen. That's another example. But almost all examples are like that. The ideas were totally created here at Caltech while the student was doing their PhD, and the student said, "I believe in what I do. It's really interesting," so they decided to become an entrepreneur and start a company. I could give you another 10 examples, but they're more or less the same.

ZIERLER: You mentioned diabetes. What other diseases does your lab focus on?

TAI: For the last 20 years, I've focused a lot on eye diseases because I collaborate with top ophthalmologists. I build devices for cataracts, for example, accommodating the intraocular lens. That has also spun off as a company. Devices that can treat glaucoma. Devices actually can treat a disease called retinitis pigmentosa in the retina where people become blind. Basic technology and materials have been developed in my lab that can be used for retinal implant to allow blind people to have partial vision again. Not ordinary vision. Also, a device that can potentially treat diabetic retinopathy. Those are the four major diseases that cause blindness. 80% of blindness is caused by these four diseases. People knew me in this field as the engineering guy who worked with that ophthalmologist to invent all these things to treat eye diseases. I also worked with a cardiologist at UCLA for heart disease. I work with many doctors. [Laugh] One graduate student's current project–in one of these brainstorming workshops, a brain surgeon was frustrated from the last 40 years of treating brain tumors. Among all cancers, the GBM actually is a special brain cancer that hasn't had much progress. Every other cancer has seen significant progress, which means the five-year survival rate's much improved. But not brain GBM tumors. I talked to him, and he was very frustrated for 40 years. I said, "Wait a minute, let's come together." We came up with a crazy idea to try to develop a new method, which is to attract and draw cancer cells to my device and then kill them.

ZIERLER: How do you do that?

TAI: [Laugh] I cannot tell you too much because it's ongoing unpublished research. But we actually think there are some proven scientific ways that have never been applied to this. It could work. The student is already trying to validate it using microfluidics right now. The graduate student is working really hard on this. Microfluidics, there's a lot of potential. This is only the beginning. In fact, I've been telling people, it doesn't matter which field I see it, "in terms of medical devices, we're still in the stone age," truly. Yet, it's a $500-billion market. Because people need it. They're desperate. CT scan and MRI, those are medical devices used every day. Then, there are other medical devices still in stone age that needs to be improved. I don't do those existing things, although AI has a huge impact on those things and engineers can have a huge impact on those fields. But then, there are future drug-delivery device. That's my courtyard. Surgical tools too.

ZIERLER: What about implants?

TAI: Implants, too, we're in the stone age. My God. I can throw a lot of examples at you. For example, I have two graduate students working on implantable microelectrodes that can be put inside a body to interface with nerves, the brain, the spinal cord, retina. Just new material, new electrodes, and that's it. It's brand new. Because the past technology gives you not-so-good electrodes. Seriously. even today, pacemakers, for example, use handmade platinum wire and use silicone to do the insulation. Those materials are not perfect. It needs to be improved. That's why I've been saying we're still in the stone age. My lab actually makes a lot of microdevices, including microelectrodes, using new materials. One spinoff is developing retinal implants to make blind people able to see. You can imagine, there are a lot of new materials we use that have never been used before.

Also, there's a lot of new computing capability because semiconductor chips have advanced so much. We've also found out that's one thing that's ridiculous if you understand. For example, semiconductor chips. Today, we're still using chips in some medical devices that were actually developed 20, 30 years ago and never renewed. Why? The FDA doesn't want it, and then the company doesn't want it. Because if you renew the chip, you have to go through the FDA approval all over again. That's because of the FDA. [Laugh] The ecosystem is very funny. In my opinion, the FDA does a lot of good things but also sort of slows down the progress (of renewing new technologies). Not like in consumer electronics, such as in cell phone communication, every month or two, there's a new thing coming out. But not in biomedicine, not for drugs, not for medical devices. It's well-known, for a new drug, it will take a billion dollars or more, 10 years or more. For medical devices, it will take maybe half a billion dollars, and even for some easy ones, $100 million and five years.

It's because the FDA is doing a really good job as a gatekeeper, making sure everything is right. But, at the same time, that means a lot of opportunities. As I said, for an advanced country and a rich society, there's no way to avoid that (high healthcare spending). People want better health, better healthcare. It's demanded. Another big factor is because we live longer. We live much longer than people 50 years ago. When you grow old, you're going to get sick no matter what. You want better healthcare technology. It's very easy to understand. When you're poor, you only think of making money. When you're rich, the first thing you think of is to live longer. If you can live longer, the next thing you think of is what?

ZIERLER: Staying healthy.

TAI: No, then you want entertainment. But to live longer, when you have money–you don't need to be rich. As long as you're financially comfortable, you want to live longer.

ZIERLER: Do you see Moore's Law at play in your lab? Is that a reality for you?

TAI: I think technology can enable that. But regulation doesn't allow that because of the FDA.

ZIERLER: Moore's Law is not as strong as FDA law.

TAI: Not for biomedicine, unfortunately. Without the FDA, yes. But then, along the way, a lot of bad things could happen. The FDA is to ensure people are protected, that no more than a one-in-a-million accident rate can happen. To get FDA approval takes a long time and a lot of investment.

ZIERLER: Do you have international collaborators, where you can test regulatory systems with perhaps less conservative practices?

TAI: Yes. Actually, it's well-known in biomedicine that, including all the major companies in US and Europe, some small countries apply loosened health laws, so it's quicker and easier to do human trials.

ZIERLER: What are some of the best countries to do that?

TAI: Mexico, Southeast Asian countries, Australia. It's not as tight as the United States. It's legal. It's just that their version of the FDA's rules are not as tight. It's not a secret. Including animal experiments. Doing animal experiments in the United States is a lot more expensive. It was known for the last 20 years, companies swamped to China to do animal experiments. 10% the cost, same results. Human trials are different. In China, it's hard. Of course, we're talking about trying technologies that have much less risk. Then, it's much easier and quicker to do it in other countries. Then, once you do that, you have very solid human trial results, you're very confident nothing can go wrong, then you go back to the United States and do the rigorous human trials. Because there's a lower risk already since you already did it in those countries.

The United States' FDA doesn't honor those results in those countries, so you have to redo it. But it's much better than doing it directly in the US, finding out it's wrong, and having to stop, after the money and time already spent. This is well-known. The trend actually is pretty bad since 20 years ago. Europe saw this problem in the United States and loosened their FDA regulations. They don't call it FDA, they call it CE Mark. It takes half the time and money to get the same thing approved in Europe for the last 20 years. A lot of companies develop the technology in the United States, take it to Europe to get approval, start selling, make money, then come back to the US.

ZIERLER: Do you see this with your students?

TAI: No. All my spinoffs, actually, try to get approval in the US.

ZIERLER: There's still something to be said for doing it here.

TAI: Yes, yes. [Laugh] What I'm doing is working on tiny medical devices, a term that never existed before, and they contain a lot of micro- and nanotechnology. It doesn't have a lot of the risks that the riskier technologies have. It can be quicker and cheaper even to do it in the United States. But that one example, device to treat type-I diabetes, it may be better to do it offshore, outside the United States, and then come back.

ZIERLER: What have been some of the advances in technology over the past 30 or 40 years that have really revolutionized things for you?

TAI: In my opinion, in terms of medical devices, the trend was accelerated in terms of available miniaturization technology, i.e., to miniaturize everything.

ZIERLER: That's different from nanotechnology?

TAI: Nanotechnology is part of it. For example, mechanical miniaturization. This is not my work, but, for example, the first pacemaker. A little electrode penetrates the skin and goes to the heart, but the wire comes out from the body and is connected to a shopping cart with electronic boxes. The patient has to pull the cart to keep the life going. And today, the smallest pacemaker is about the size of a quarter, and the whole implant is below the skin. That's gone through many generations of miniaturization. It's mechanical miniaturization with electronic miniaturization. Same thing has happened to so many other medical devices. In the past, they'd draw the blood, use a Petri dish, flasks, tubes, just like every high school student learned to do in the wet lab. Now, just one drop of the blood on the sensor, and you're done. Then, you just throw everything away. The science never changed. It's just miniaturization of fluid handling, the size, mechanical design. But there's also new nanotechnology in there, new molecules and all that for new tests. But nothing has changed. The science never changed. Really, just the size has changed. Miniaturization changes medicine so much. That's the opportunity. And then, I happened to do research on small devices and how to miniaturize things. I'm one of the lucky ones. [Laugh]

ZIERLER: What about computation, either for analysis, or for modeling? What advances in computation have been important for you?

TAI: To me, not much.

ZIERLER: You don't do much modeling or simulation?

TAI: I do. But in my opinion, I'm not good at that. Especially the AI thing. If you don't talk about AI, computer chips and all that enable miniaturization because it used to be a big circuit board, but now it's a tiny chip. I have a project to make a pacemaker. We know today's pacemaker, there's a metal case that big, and then the wire goes into the heart. We want to make a pacemaker one centimeter long, two millimeters in diameter, a tiny tube. Then, you put it on the surface of the heart, and that will do the pacing. Smaller is better. Your body will ignore it. It can be catheter-delivered. There's no big wound. And it'll do exactly the same job. From a shopping cart of electronics to a tiny, little chip that does computation on it. But unfortunately, to shrink a large chip to a tiny chip is not a PhD to me. What I actually care about is a medical device being shrunken to the size of a grain of rice. There are a lot of new things there. New design, new mechanical principles you need to deal with. That's a lot of intelligence required. That's a PhD to me. For example, my students also use AI, but not as a key part of their thesis.

ZIERLER: It just makes it easier?

TAI: Makes it easier, better. But my research deals mostly with physics, physical sciences. That's what I enjoy most.

ZIERLER: Who do you collaborate with on campus, even beyond EAS?

TAI: I collaborate with Professor Emami. She's the one who reduced electronics to a tiny chip. Her students can get a PhD from that. Mine don't. I used to work with Richard Anderson, who does neural engineering and developed a tiny flexible electrode to put inside a monkey's brain. He's a scientist. He wants to use those ideal devices, but he doesn't know how to build them. I collaborated with him. But in my domain, I truly think miniaturized medical devices has a great future. And I've been doing that for 30 years.

ZIERLER: And that's where things are headed.

TAI: Yes, continue to go into that direction. For example, if you want to treat a brain tumor, I need to build a device and put it inside the brain. It has to be small. Or to put a device inside the eye. The eye is not that big. I want to make a pacemaker the size of a hair.

ZIERLER: Have there been advances in microscopy so that you can see what you're doing?

TAI: Oh, yes, from my colleagues. Microscopy or imaging technology that can allow you to see and diagnose better. Many of my colleagues are doing that. They call it medical optics. It's amazing. There's huge advancement in that also benefitting from new technology for miniaturization. They want to make tiny, little microscopes. A microscope is pretty big. I say, "No, no, no, my colleagues make tiny chips that can do the same imaging as a microscope." Miniaturization is changing the world. The old ways of radio is a big box a soldier carried. Today, it's a cell phone. In fact, the radio is a very small part of the cell phone. And there are actually people who say, "I don't need radio." We call people through internet now, voice-over-IP. But everything else is still miniaturized. People want smaller, better, cheaper. I also think miniaturization has a huge advantage why it's so important and impactful. It's manufacturing. When you build small things, you never just build one at a time. 100 years ago, we built things one at a time.

Even today, when you build a car, you build one car at a time. One production line can give you 10 cars a day. If you want 100, you have 10 production lines. Semiconductors are different. Semiconductors count the surface area as real estate because the device is so small, and your total area's big. Even if you make only one chip, still, there are 9,999 chips right there. [Laugh] Mass production when you reach miniaturization. All the devices I'm talking about are moving toward that direction. You don't just miniaturize them, you also mass produce them at low cost. You make it popular. Everybody can afford it. I also see that as a big problem for US healthcare. Cost is so high. The US actually spends more than $8,000 per person per year on healthcare, yet life expectancy is no more than that of Japan. Japan spends only $3,000 per person per year on healthcare. But health costs actually continue to go up.

ZIERLER: This is about genetics and lifestyle also, though.

TAI: A lot of factors. But one thing we do know for sure is the cost of medical devices is high. If we are able to make semiconductor chips, where each chip is 10-cents, the machine you put together has to be low-cost. That's why a cell phone is ever more powerful, yet the price is affordable. And I think we need that for medicine, too. In the future, a lot of devices people will use–for example, complete blood count for $1. Today, insurance charges $50. It could just be $1 or cheaper. Low cost, to me, is very, very important for future medicine. No doubt about it.

ZIERLER: And miniaturization is part of that.

TAI: It's important technology to support. If you can miniaturize and combine that with mass production, you can get there.

ZIERLER: Last question for today. You say we're in the stone age right now with all of these devices. How do you see your lab in the long-term helping to get the entire industry out of the stone ages?

TAI: It's not just me. In my class, I tell the young people, "This field needs you guys. Science says it can be done, but it will continue to take a lot of effort." The good thing is, I see that, I believe in that. There are always a lot of smart people looking for opportunities. If a good opportunity presents itself, smart people are going to come. That's how it works, really. I think this field has already drawn so many people into microfluidics, micromedicine, lab-on-a-chip. I see a lot of people jumping into this direction, and it's because this direction is so attractive. It makes sense.

ZIERLER: You're excited.

TAI: We just need to spread the news. We just need to bring this information to smart, young people. They will join us.

ZIERLER: They'll take it from there.

TAI: That's right. Smart people can smell opportunity, and they will jump into this area. That's what I believe. I say a lot about future medicine. People want cheaper and better healthcare. The whole Earth is spending more and more money, not just the United States. This is a great industry, great direction to work in. If the opportunity's obviously there, people are going to jump in. This is what I see everywhere, every country. You go to conferences, and they're growing. You see young faces. That's always a good sign. [Laugh] That's the way it is. It's because this field and direction is that attractive.

ZIERLER: On that note, next time, we'll go all the way back to mainland China, Taiwan, and ultimately, the United States.

[End of Recording]

ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Thursday, March 16, 2023. It's great to be back with Professor Yu-Chong Tai. YC, great to be with you again. Thank you so much.

TAI: Oh, thank you, David. I had an excellent time in our last meeting.

Family Roots from Mainland China to Taiwan

ZIERLER: Let the fun continue. In our first conversation, we did a great tour of all of your approaches to the research and engineering. Let's do some family history now. Let's go back to mainland China, and Taiwan. Why don't we start with your grandparents on both sides? Tell me about what you know of them and if you had relationships with any of them.

TAI: I never met my grandparents because my father and my mother came from China and went to Taiwan with Chiang Kai-shek because Chiang Kai-shek lost a war to the Chinese communists.

ZIERLER: This was 1949.

TAI: This was 1949. Actually, that was exactly the year both my father and mother came to Taiwan. Of course, they didn't know each other there. It was more like 10 years later, they got married in Taiwan. I was born in Taiwan. I grew up in Taiwan. After I did my military duties–at that time, every man after college had to serve two years of military duties, like ROTC. Then, in 1983, I came to the United States. I attended UC Berkeley for my PhD.

ZIERLER: Where in China were your parents from? What regions?

TAI: Actually, my grandparents and my father were in a small town, a rural area in Anhui province. They were basically farmers. It's a small place, small town. But that town, and the whole area, is known for farming. My grandparents were farmers. My father, at least to my knowledge, was a college graduate. He was one of the rare few people who actually completed a college degree. My mother didn't even finish middle school. that was because the war happened. My mother's education was sort of completely interrupted.

ZIERLER: Were your parents' families under threat from the Japanese occupation?

TAI: No. Well, no, in the sense that they always tried to escape, and both of them were successful in avoiding the Japanese-occupied area. But yes, because they told me so many stories of how they managed to avoid the Japanese armies. For example, my mother told me when she was in middle school, it was never a school. Basically, the whole school was led by the principal. Everybody carried some bread and water and just ran around to actually avoid the war. That was actually how my mother finished middle school. My father was much older, so he'd actually already finished college. In Chinese, we say the KMT, Kuomintang. He was with the government. He wasn't a soldier, just a government employee. My father actually had a better life than my mother. [Laugh] They were affected by the war, but then they never stayed in the Japanese-occupied area.

ZIERLER: When they moved to Taiwan, did they ever have contact with their parents again?

TAI: My father actually sporadically told me that he heard some news from my grandparents. But my own memory is that for at least 20-some years, there was absolutely no communication. Because even letters couldn't reach the other side for 20-some years under Chiang Kai-shek and his reign in Taiwan. There was no phone calls, no letters. Actually, people understood they shouldn't do it, especially my father, who was with the KMT. Writing letters to the other side was a no-no. [Laugh] Not mentioning trying to make a call. But I know people flew to Hong Kong or other places to try to reach out to their families.

ZIERLER: Your father was part of KMT. I understand that's why he would've gotten to Taiwan. What about your mom? What was her story?

TAI: Well, my mom's father actually was also KMT. He was able to take my mother to Taiwan, too.

ZIERLER: Did he stay?

TAI: He stayed, yeah.

ZIERLER: Did you know him?

TAI: I knew him very well. Actually, when I was little, almost every summer, I would go to my mother's house. I remember there was no good transportation, only buses, to travel to my grandpa's house. It was basically in the middle of farmland. It would take maybe six hours, three buses.

ZIERLER: Did your father stay in government in Taiwan?

TAI: Yes, all his life. He's actually an engineer. He was with the Department of Economy and specialized in water utility.

ZIERLER: So like a civil engineer.

TAI: Like a civil engineer but really focused on water utilities, like river management, dams. Actually, Taiwan built a couple of small dams, and he participated in that. He was also, for a period of 10 years, on a team to Africa. At that time, Taiwan actually sent farming teams to Africa to teach the ally countries how to farm. And water is a very important part of that, so my father was put on this–they called it a support team to Africa. I remember, for almost 10 years, I didn't see my father. That was my younger days.

ZIERLER: Where did your parents meet?

TAI: In Taiwan. Through introduction. Because there were a lot of single men and a lot of single girls at that time due to the war. They were introduced to each other, and they got married.

ZIERLER: Where in Taiwan were you born?

TAI: Taipei.

ZIERLER: Did you grow up there?

TAI: I actually grew up on the border of Taipei. At that time, my family wasn't rich. We belonged to the middle class, but the lower half of the middle class. We didn't face hunger, but…

ZIERLER: No fancy vacations.

TAI: No. I couldn't recall any vacations at all. Except every summer, my parents actually would send the kids, including me, sometimes my brother, to my grandpa's house for the whole summer. [Laugh] I guess they didn't need to take care of me that way. That was it, there was no vacation. I remember the first vacation was after my graduation of middle school. this was a school trip, all kids were welcome to join. The school arranged a graduation trip, so I got a chance to visit the middle and southern part of Taiwan. I remember very clearly, that was the middle school graduation trip. That was my first away-from-home trip if you don't count my trips to my grandpa's house.

ZIERLER: Your family lived in a house, an apartment?

TAI: In those days, because my father was a government employee, they all stayed in government dorms. It's not like today where you build a skyscraper, and then you have apartment A, B, C, D. Everybody actually sort of had a "house." The quality was questionable. But it's very hard to have two- or three-floor buildings. At that time, only very rich people had them. The government actually assigned sort of dorms to their employees, and I remember we lived in a house very close to the river. Believe it or not, when I was young, there were floods. [Laugh] I went through so many floods. The water actually came up to my waist. Many times, you woke up in the middle of the night, and it was still dark, like, "What happened? It's so wet. [Laugh] Run, run." I grew up like that. But I had a lot of fun. I have a lot of memories. There was no toilet. Everybody lived like that in my area. It was a government dorm area.

ZIERLER: Tell me about the schools you went to growing up.

TAI: The first grade school was an elementary school. It was very small, only one class. It was a school started and designed for people living in the whole dorm area, so not that many kids. I was little, about 4 or 5 years old when I started. The way they arranged that, they knew which kids were in which families. They organized the group walk. From house one, house two, house three…, they gathered kids going to the school. We'd gather, and there was a leader, an older kid who'd pick up the kids, and we'd walk together. Half an hour, every day, we'd walk to and from school. It was a small grade school. I wasn't that great of a kid, but my mother did one thing. My mother totally believed in education. Not my father. But my mother was willing to spend all the money to send kids to a good school if there was a chance.

ZIERLER: Opportunities that she didn't have.

TAI: Yes. When I was in fifth grade, my mother managed–this is amazing and I don't know how it happened–to send me to a better private school that was pretty far away from my original house. But I wasn't officially admitted into that school. But at that time, because it was a special situation in Taiwan, the government allowed schools sometimes to take kids temporarily for whatever reason, people were moving around. It's like a transient student. I was a transient student, I joined this private school, and I was able to sit in the class and had my little desk. One class was typically 50 kids. That was the norm. I was number 51. [Laugh] In the very back, there was a tiny, little wooden desk and wooden chair. I was arranged as a transient student to study in the private school. there's one thing I'd like to talk about because to a lot of people, it's a painful thing, but to me, it opened up the only opportunity for my life.

When I was in the sixth grade, in those days, when you reached the age to go to middle school from elementary school, you had to take tests. And then, from middle school to high school, you had to take tests. From high school to college, you had to take tests. GPA didn't matter at all. It was the tests on that one day. For high school and college, it's two days. But from sixth grade to seventh grade, it was a test, one day. I did just barely enough to stay in the same school, so I became a formal student in the middle school. Then, when I was in ninth grade, there was another test called a national high school entrance exam. Two-day test, five topics, including math, Chinese history, English, etc. I did good enough to go to the "best" high school in Taiwan. Then, after three years of high school, there was a two-day test with six topics. Because I chose to go the natural science route, so including physics, chemistry, English, Chinese history, math. Again, I did well enough to go to the best college, NTU, in Taiwan.

ZIERLER: National Taiwan University.

TAI: That's right. The number-one department in the natural science route, electrical engineering in National Taiwan University. If it weren't for this test system, I don't know where I would've ended up. [Laugh] GPA didn't matter? It's a one- or two-day test. You do well, you get into a better school.

ZIERLER: Did your father being an engineer influence you? Did you appreciate what he did?

TAI: Not at all. As I said, there was 10 years' time when I was in grade school that my father was in Africa. I guess it helped the family. The pay was better. At least my family didn't have money issues. That, I really appreciated. I didn't have a lot of memories with my father at all. He was mostly away from home. There was no baseball time, no basketball time. Zero. I do know when I was really little, my father was in the house, then there was a long period of time, and when he came back, I was already a teenager. [Laugh] It was pretty much my mother who took care of the family. But my mother, I told you, barely graduated from middle school. she was a really tough lady who did everything she could to send my brother and sister to the best schools she could.

At one time, I told my mom, "I'm happy to stay in the public school. It's much cheaper." I saw all the money she saved for the year went to school because private school tuition is very expensive. I told my mom, "That's okay. I'm happy to stay in the public school." But after middle school, I was able to take advantage of the testing system. I went to all the public schools, high school and college. That didn't add any burden to my family anymore because the public school is much cheaper, and yet, you can get the best education. In a way, I had to work hard not to drop out of school. [Laugh] My older brother and older sister didn't really make it through to college. I was the only one who went to college.

ZIERLER: Growing up, you learned English?

TAI: I learned a little bit of English in middle school. I actually remember, I learned the word "apple" in middle school. You learned some vocabulary, but not much more than that. In high school, yes. In high school, there's a formal English class that's pretty intense. Because in the national entrance exam from high school to college, English was a must-take topic. It was a mandatory topic, you had to test on it. High school, we learned English a lot more. Then, in college, we also had two years of English class, including colloquial English. But then, I didn't know how important English was. [Laugh] But I did know one thing. Somehow, since elementary school, my math was pretty good. It helped me through my education so much, even when I became the transient student. All other topics, geography, Chinese, I was so horrible. But my math always stood out. [Laugh] I went through this whole education leveraging, I guess, what Caltech likes, math, physics, and chemistry. But English was a must-take exam topic, so I studied some English. I wasn't that great. But then, I realized English was super important when I attended UC Berkeley. I had no problem taking care of the technical courses, but I realized English was so important, so I guess I paid a little bit more important to English. I really learned English after I came to the United States.

ZIERLER: Growing up, did you feel the threat of a Chinese attack on Taiwan?

TAI: No.

ZIERLER: That was never a concern?

TAI: When you get older, sometimes some ideas just stop popping up. I realized one thing. We knew there was a threat. Everybody was talking about the threat. Who knew what would happen? But I realized that if it was coming, there was nothing I could do. [Laugh] If the war really happened, there was nothing I could do as a teenager in Taiwan. If there's nothing you can do…

ZIERLER: Just go on with your life.

TAI: Yeah, just focus on your life. That's what I thought. And actually, I was quite interested in learning in school. All my life, actually, I really enjoyed learning new things, so that's what I did. I focused on learning. And I had a great chance because I was in the best high school and college in Taiwan. Actually, when I was in college, nobody talked about an attack anymore because China, at that time, was a very closed society. And then, Taiwan actually was more open. In my day, take a look at my classmates. About 70 to 80% went abroad for higher education. I was one of them. I learned in that kind of environment, the United States actually has the best graduate schools. We learned that you can learn a lot more. Honestly, I didn't learn much in college. I came to that conclusion only after I went to Berkeley. When I went to Berkeley, I started taking courses and talking to people, trying to understand what would be interesting knowledge I could learn.

Then, I realized my college was like a kindergarten. [Laugh] That was my feeling for my first months in Berkeley. "Oh my God, that is true knowledge. They're at the forefront of knowledge compared to what I learned in college." Then, I realized one thing. I did benefit from my college education in terms of fundamentals. The school still taught math, physics. And then, the college teachers, I guess in those days, all they could do was teach the fundamentals better. That helped me so much. When I went to Berkeley, it was a dream come true, really. They'd go through something very detailed, talk about how an electron behaves when it goes through an energy field, you can do a calculation and figure out the energy, its quantized energy state, then actually use that to build devices, transistors. It's amazing. Just a different level. It's not just one level of separation. One is close to heaven. One just prepares you to learn more.

From Taipei to Berkeley

ZIERLER: At college, did you live on campus or commute from home?

TAI: I actually left home after high school, for a few reasons. For one thing, my sister actually got very sick with lupus. At that time, it was regarded as an incurable disease. The family had a lot of pressure. I was the only student in my class that applied to stay in the dorm. At that time, my college dorm was only reserved for people who actually didn't live in Taipei City. My home address was within Taipei City, so I wasn't qualified. But I sent in my application to the school saying, "If there's any empty space, I would like to stay on campus." Part of it was so I could focus on my education. In fact, my mom told me to do so. She's unbelievable, come to think about it.

Anyway, I was the only Taipei student living in the dorms. At that time, the dorms–I know my three kids went to college, and when I took them to the dorms, I said, "Wow, this is not a dorm. This is a grand hotel." [Laugh] The dorm I stayed in was a tiny, little room with four bunk beds, eight students. Everybody had a tiny desk, a tiny wooden chair, and that filled up 90% of the space. That was the dorm. Then, you had a shared shower room, shared hygiene room. But inside the dorm room, eight people, four bunk beds next to each other. Little rectangular room, two bunk beds on this side, two bunk beds on that side. In the middle of the bunk beds is eight desks, four on one side, four on the other. If you wanted to walk out, you had to say, "Excuse me," and squeeze out of the little space.

ZIERLER: What did you study in college?

TAI: Electrical engineering. At that time, in Taiwan, you didn't get a lot of information about your future, what you could do. It was an exam. One test, you're done. Because of the test system, all departments had been ranked, which department students wanted to go to most. The system actually is this. You take the test, and then you get your score. The highest-score student chose which departments they wanted to go to first. After number one, then number two. Immediately, there's a ranking of departments students want to follow. Electrical engineering, at that time, was number one. That was the time semiconductors started to boom, in the late 70s.

ZIERLER: This is the beginning of the semiconductor industry in Taiwan.

TAI: Yes. I remember when I was a teenager, the first wave of the manufacturing industry was to make a discrete-component simple PC board. At the factory, I saw people making capacitors and other tiny little things. They made small PC boards. It started when I was a teenager. But there was no semiconductor manufacturing, it was only assembly.

ZIERLER: And the raw materials came from elsewhere.

TAI: That's right, from the United States or elsewhere. But when I was in college, that was the time that microprocessors started to show up, which meant that people could have personal calculators. Oh my God, calculators started showing up when I was a college kid. They changed your life. Compared to hand calculation and sliding rules, electronic calculators started to show up when I was in college.

ZIERLER: What about computers? Were there any computers in college?

TAI: No, actually, there were microprocessors. You had to use what we called machine language, the lowest level machine language programming. When I was in college, there was a class to teach computer language, and we learned Fortran, which was the first computer language. There were no personal computers, but there was a big, giant server. It was called UNIVAC. I saw that in a movie. It was a vacuum-tube computer, American UNIVAC computer. The government spent a lot of money to put that computer up there. We had to write programs, learn Fortran language. Then, we had the punch cards. This IBM machine would punch holes on a paper card. Then, whatever you wrote, you ended up with a deck of cards. The cards were pretty big. If there was a hole, it was a zero, and if there was no hole, it was a one. Whatever little program you wrote, it was a thick deck. I hated that experience, so I decided not to go in the direction of computer science.

Because then, we figured out we'd write a program, first of all, you'd have to write it down on a piece of paper, then you'd actually have to go to the IBM card-punching machine. There was always a long line. I'd go there and wait an hour, then your turn came, you punched each individual card, then you had a deck of cards. Then, there was a program submission room with a worker, and they'd take your deck and make a record, then put it into the IBM card reader. Then, you could come back a day or two later to get the results, then you'd run it on the computer. Then, a day or two later, I'd go back and pick up my result as a sheet of paper from a dot-matrix printer. They'd tell me, "Oh, you had a typo on card two and card four. We couldn't read it. We didn't know what it meant." Are you kidding me? [Laugh] For example, instead of typing "then," I typed "thn."

You'd wait in line for the punch cards again to correct the error, punch a new card. Then, you go back and redo this process, and two days later, you got a printed sheet of paper saying, "On cards 9 and 10, you had a typo." "Why didn't you tell me about the errors?" The computer just couldn't do it. If it read two or three errors, it would stop. To correct a typo–not typo, I shouldn't say that. It's a card-punching error. A simple program may take you one month to run. I'm not that kind of person. I found it totally a waste of time. I can spend 10 hours thinking about a program. That's fun. I actually ended up doing this matrix inversion. If you input a matrix, I'll tell you the inverted matrix.

ZIERLER: What does matrix inversion mean?

TAI: If there's a matrix, you can find another matrix–matrix A times matrix B then equal to matrix one. All the diagonal items become one. We call this an inverted matrix. A times A - 1 is equal to 1. if you have matrix A, the software I wrote can give you A - 1. It's very important in linear algebra, especially today because all the AI computation, you do a lot of matrix inversion. I wrote a simple program because I'd learned some mathematical algorithm to find the inverted matrix of a matrix from a textbook. It wasn't invented by me. [Laugh] Just going through that class, I suffered so much. I said, "The first few hours is interesting when you think about mathematics, how to invert a matrix. But then, punching cards, the computer telling you there's a typo, redoing it. If that's computer science, I'm out." Maybe I made a mistake, but I decided to stay within electrical engineering, the hardware side, devices, transistors, electromagnetic fields.

ZIERLER: Were some of the professors in electrical engineering involved in the building up of the semiconductor industry in Taiwan?

TAI: Not that I know of. Actually, semiconductor fabrication was introduced into Taiwan after I left. That's when TSMC was started. That was exactly the time I went to Berkeley.

ZIERLER: When did you get the idea or encouragement that you could go to graduate school beyond Taiwan?

TAI: Actually, in my day, most of my classmates, and actually the whole electrical engineering department, after graduation, came to the United States.

ZIERLER: Only the top students, not everybody?

TAI: Believe it or not, 70% came to the United States. They scattered around to different schools. Berkeley, at that time, electrical engineering was ranked number one. It was very crowded. In fact, Stanford didn't rank higher than Berkeley. Berkeley was number one then.

ZIERLER: Did you think about Caltech?

TAI: I didn't know Caltech. I didn't apply to Caltech. My goodness. [Laugh] I have a lot of stories about Caltech. In my day, UC Berkeley was ranked number one for EE. MIT, Princeton, Stanford. I only applied to six schools. Those four, plus UT Austin. I forgot number six, my backup school. Why UT Austin? It ranked pretty well and was very cheap. I really couldn't afford MIT, Princeton, or Stanford. Berkeley's tuition was much lower, but they were ranked number one. And I got in. I got a fellowship, so I decided to go to Berkeley.

ZIERLER: Was your first time on a plane when you flew to the United States?

TAI: Yeah, that was my first time.

ZIERLER: What was that like for you?

TAI: It was amazing, my goodness. The whole trip was quite an experience. I actually had classmates already at Berkeley. In fact, I had one classmate who didn't need to serve military duty.

ZIERLER: Yeah, when did you serve, before or after college?

TAI: After. Between 1981 and 1983.

ZIERLER: And this is full-time service?

TAI: Oh, yeah. You still in the military unit. You can't get out.

ZIERLER: Did you defer your acceptance, or did you wait to apply until after?

TAI: No, I actually applied a year before admission, but I had to apply for schools while I was in the Taiwanese Air Force.

ZIERLER: Did you fly?

TAI: No, no, no. [Laugh] I didn't fly. I couldn't fly. But there's another national test. Because I graduated from NTU in electrical engineering, I took a test on radio electronics. I guess I did well again, so I was awarded–if you test well, you're awarded as an officer. If you couldn't do well, then you are G.I. Joes. [Laugh] I was promoted to the rank of officer, second lieutenant. Actually, there's air-force electronics. That means electronics used by the Air Force. I became an officer of the Air Force. For almost two years, my job was to maintain the radios in the American-built airport in the middle of Taiwan. In fact, I was in charge of fixing and repairing whatever radios on the Air Force base. It was a fun two years. Really fun. In fact, it's ridiculous, the first day I reported to the airport, they said, "Okay, you're an electronics officer. There's a radio station. Go report to the officer in the radio station." I went in there, half of the radios were broken. Vacuum-tube radios. Because the fighter planes have to communicate a lot. You have to keep the radio communication open. It's a whole building with vacuum-tube radios, one after the other, almost half of them broken. [Laugh] I went there, and in one week, I repaired all of them. Just suddenly I was in charge of them. [Laugh] That's what I did in the Air Force.

ZIERLER: You had a good time.

TAI: I had a good time.

ZIERLER: Did you ever think about staying, making a career in the military?

TAI: No. Hell no. I knew I could learn a lot more. I'll tell you, the Air Force guy wanted me to stay. "Sure, if I can't get any school to admit me." I applied to those six schools, and I got admission from Berkeley with a scholarship. I also got accepted for Princeton. I was waiting for Stanford. I didn't get admitted by MIT [Laugh] but Berkeley was ranked number one in electrical engineering. I had a good time doing my PhD. And after Berkeley, I came to Caltech.

ZIERLER: Tell me about your first few days at Berkeley. What do you remember?

TAI: I have a story I should tell you. Because I already had classmates at Berkeley. They went before I did because I had to wait until I finished my military duties. I had classmates who didn't need to do that who were there. Email was there already, so I made a communication, "Please pick me up. I'm coming on this day, this flight." I realized actually, my first story is, I came off the airplane, picked up my luggage, and nobody was picking me up. In the middle of the San Francisco airport, nobody was picking me up. There was no cell phone at that time. I just called a taxi. I called a taxi, showed the address. Bad English. I went to the apartment all by myself. That was the first lesson I learned, you can only count on yourself. Don't count on anybody else. That's absolutely the first lesson. The first flight. [Laugh] I was left alone in the airport with the language barrier. I told myself, "Count on yourself. Do everything by yourself unless necessary." It turned out the guy totally forgot. That's okay. And then, when school started, the first few days, I was in the apartment I rented with three other guys. Four students jammed in a not-so-big apartment. Very cheap.

ZIERLER: But you were used to this.

TAI: But then, I learned, actually, my other friends' apartment was even worse. Even cheaper. So bad, they couldn't tell us the original color of the carpet. It had not been cleaned for maybe three or four decades. It was just black and dirty. Anyway, cheap. And then, when school started, that was a cultural shock. I attended classes, and of course, it was all about semiconductors in the beginning. Semiconductors, circuits, transistors, and all that. It's amazing. That's the time I realized it was heaven. You could learn so much from the best people in the world. Seriously. All teachers are top guys in the field of electronics and transistors.

ZIERLER: Did Berkeley feel big to you?

TAI: No, because National Taiwan University is about the same size. Lots of people, lots of students. Physically, it was not big to me. It was the quality, teachers and professors and how much they knew. They were all leaders in their fields, every course. I wished I could've taken all the courses. Of course, I realized pushing forward was important, research was important. When you did enough, you got the PhD, then you're on your own. That was also the time Silicon Valley started to take off, because the semiconductor really took off. A lot of chips were available. Apple started at that time, Microsoft started at that time when I was in Berkeley. The early 80s. Unbelievable. Intel, AMD, they were all there. That's my education.

ZIERLER: What did you do for the summers? Did you go home, stay on lab?

TAI: I stayed on campus doing research. My life at Berkeley was quite simple. Take courses, learn as much as you can. Do research, and do as much as you can. In fact, when I was a junior or senior, starting the third or fourth year, I was one of the very few students who would never leave campus, day or night. I remember, there were nights at 10 pm, three of us would happen to get together outside of the lab. And we'd say, "Do you want a break?" That was actually Saturday night. "Do you want a break?" "Yeah, sure." Went to pizza, went to a movie theater. That was it. That was our life. Once in a while, we went crazy. I remember, we asked each other, "Which movie should we watch?" And the other guy said, "I want violent and bloody movies." [Laugh] We decided to watch the movie RoboCop. It's violent and a lot of blood. That tells you we'd have been working days and nights. [Laugh]

ZIERLER: Who ended up being your thesis advisor at Berkeley?

TAI: It was Professor Muller. Unbelievable guy.

ZIERLER: What was his field? What did he do?

TAI: It was funny how I chose the field. At that time, when we'd just entered Berkeley, Berkeley had a system where students came, and then it became kind of hectic when the students wanted to look for professors. Professors were picking the students. That was exactly the time I was knocking on professors' doors, saying, "Do you have an opening? Maybe you would be interested in me." Well, the first professor I asked couldn't take me because a student who'd arrived at Berkeley much earlier had already talked to him, and he'd already made a decision. He had no opening. He ended up being a very famous guy who invented the modern transistor at Berkeley. I looked at others. Then, I talked to my thesis advisor, Professor Richard Muller. He still had an opening, so I joined his group. But he doesn't do much with transistors and all that. He does sensors and actuators. He made his name in sensors.

ZIERLER: For what kinds of applications?

TAI: At that time his students were making accelerometers, pressure sensors and so on. I didn't go the route of integrated circuit design or new transistor design, so I ended up in the direction of making sensors. But I guess something happened. It turned out, making transistors, including the design of transistors, learning about transistors and how to make transistors, can be totally applied to making sensors. In fact, it became a revolution. We don't have to talk about large sensors anymore, we can talk about microsensors with transistors. That changed everything. We call it smart sensors.

ZIERLER: This is your first exposure to what we would call nanotechnology?

TAI: Well, at that time, there was no such term as nanotechnology. It was called microtechnology, or micromachines, or MEMs. DARPA started the MEMs program in early 90's, and they had money, so everybody used the term MEMs. Before that we called what we did "micromachines." That actually was what I did, make micro gears, springs, joints. And then, I made the first working micro motors.

ZIERLER: Did you learn integrated circuits? Had you heard about VLSI?

TAI: Oh, of course. Berkeley was the best for that. I took a lot of classes on VLSI. In fact, before that, it was LSI, large-scale integrated circuit. Then, it became VLSI.

ZIERLER: In Taiwan, you'd never heard of Caltech. When you were at Berkeley and learning about VLSI, did you learn about Carver Mead?

TAI: Yes, I did. I learned about Carver Mead. But to me, Caltech was a small school, top-graded, did things its own way. Because I went to conferences and all that in my field, and I just didn't see students from Caltech. [Laugh] I encountered students from MIT, Stanford, a lot of other schools. But no one from Caltech. Zero. Actually, I have a story to tell. At that time I graduated from Berkeley, I was looking for jobs. The first school I applied for a job at was Stanford. Put a package together, sent it in, did an interview, gave a talk, and then I waited. Of course, I got whispers that Stanford was interested in me and wanted to give me an offer. I said, "Oh, that's nice." That was it, the only school I applied for because it was right next to me. People said, "They're looking for people. Send in your package." I said, "Oh, okay." I also submitted my package to Bell Labs. Bell Labs was interested in what we did, micromachines. I went there and interviewed. In fact, I graduated in 1989, and that was a really bad year. Bad economy.

I only interviewed at these two places because I was told they were both going to give me an offer. One day, I went to another professor's lab with my friend. I saw that one yellowish paper. It's a letter-sized paper, yellowish, been there for many years, more than two years, probably already three years. "Caltech electrical engineering faculty opening." Then, I see the phone number. I see, actually, Professor Dave Rutledge. "Blah, blah, blah, we have an opening. If you're interested, please contact Professor Dave Rutledge." I put down the number and gave him a call. [Laugh] I said, "Hey, Professor Rutledge." "I'm sorry, it's kind of late. That's been posted for a long time." "But is the opening still there?" He said, "Yeah, the opening is still there." "Then, can I send you my package? I just graduated." He said, "Of course. Everybody's welcome." I sent him the package that I made for Stanford and Bell Labs.

VLSI and Micromachines before Nanotechnology

ZIERLER: We have to go back now. We have to know what's in the package. What did you do for your thesis?

TAI: I did these micromachines. We learned a lot from VLSI circuits, how they built transistors, to build micromechanical parts.

ZIERLER: For what? What were the applications?

TAI: Oh, God. There are so many we can talk about. In fact, the story is this. I first came up with the idea because at Berkeley, there was a visitor from the Henry Diamond Lab, which was affiliated with the US Army. They actually came to give a talk and try to, I guess, recruit students. I went to the talk, and then I talked to the guy. He said, "We're actually making microsensors." "What do you think it could be used for?" The guy said, "Hm. Detonators." It came from his mouth. He said, "Small and precise. We can use microsensors for detonators." I came back and said, "Detonators?" It triggered my mind. Then, I went to the library to study detonators. Several hundred years ago, they were mechanical latches. When the vibration happened, the latch broke, and the bomb went off. When the bomb hit the ground, the mechanical latch opened up.

I said, "Wow, detonators. How would I build detonators? Can I actually use semiconductor technology to make the existing detonator, an older mechanical structure, into a tiny, little mechanical detonator?" Natural. I didn't build a transistor; I used the same technology to build that. It went off. I started to actually build the gears, the latch, all that. But it at least triggered the idea that, "Oh my goodness, you can actually build all kinds of mechanical parts." What do you use them for? There are so many examples in the real world. At that time, we talked about watches. Swiss watches. All the micro gears are handmade, so that's why they're so expensive. "We can use semiconductor technology to make that." In fact, one company's already doing it. They use exactly the semiconductor technology, what we call micromachining technology, to build parts to put in a watch. It's commercialized by one Swiss company. There are many other possible applications.

ZIERLER: We talked about computers at National Taiwan University, or the lack thereof. What about computers at Berkeley?

TAI: Oh, computers at Berkeley were very advanced. I had many friends in computer science. Because in college, I went through this ordeal of writing and running programs, I decided to go into physical electrical engineering. At Berkeley, although they call it the Department of EECS, in fact, it runs like two departments. They're very far away. I stayed within the physical electrical engineering side. I'd already made up my mind. And actually, working on micromechanics opened a big door. I didn't hesitate to continue that route because I truly believed in it. We actually called them microtransducers. Sensors, an actuator, a micro motor. The actuator would give you some force. It was relatively unexplored. There was no such thing, so a new field was born. I put a lot of examples of small things I built in my package, and I sent it to Professor Rutledge. Amazingly, in three days, he called me back and said he wanted to invite me for an interview. I said, "Sure, as soon as possible." [Laugh] I came down three days later. Within one week, it was all decided and done. I came here for an interview, I gave a talk, I talked to many faculty–the interview process at Caltech is amazing. They filled out a schedule - every half an hour you meet and talk to someone. It was a very busy interview.

ZIERLER: Because there were all of these industrial applications for your research, did you ever think about going into industry? Did you get recruited in industry?

TAI: It was a new field and the thesis I did is very new. There was no industry working on that. And I thought there was no doubt the field should continue to move forward, and the best place I could go was into research or a university.

ZIERLER: Were you thinking about biology at all in graduate school?

TAI: No, I knew zero biology.

ZIERLER: Were you aware there was a field of bioengineering?

TAI: No. Because in Taiwan, there was no bioengineering. Biology, in the direction of bioengineering or medical engineering, is maybe 50 years behind. In the United States, people were already moving forward. We call it biology, but there are new branches into molecular biology. The US already moved forward with molecular biology. In Taiwan, for example, people still were only doing cell biology. That was the frontier. But in the United States, molecular biology was already almost done, right? Or biochemical biology. Biochemistry was part of it. Biomolecules. They were already talking about DNA, RNA, proteins. In Taiwan, I never heard about that. In my education, I never heard about it. I only learned so little about biology in high school because there was, one year, a mandatory course in biology. We talked about how the lungs worked. Very fundamental stuff. That was my impression of biology. [Laugh] I knew zero biology, really. [Laugh]

ZIERLER: Tell me about your interview experience at Caltech. First of all, when you got here, what was it like to be on campus?

TAI: It was amazing. I remember Professor Rutledge was so nice. He actually booked the Einstein Room for me. I stayed in the Einstein Room the first visit. I learned later it was the Einstein Room. Wow. But then, the room wasn't really that great but with antique furniture. It wasn't as comfortable. I thought the Holiday Inn was better. [Laugh] Later on, I learned, "Wow, there's heritage here." It was prestigious to stay in there. Aside from that, I was amazed.

ZIERLER: Who do you remember meeting? Did you meet Carver on that first trip?

TAI: No, I didn't meet Carver. I met mostly electrical engineering people. Professor Papas. He was over 100 years old at that time. I talked to him, and he told me a lot of histories. He said, "My research, I only need a piece of paper and a pencil." When he was young, he actually worked with some famous Chinese guy during the Civil War. I had a great talk with him. Of course, I met Professor Middlebrook. He was executive officer of electrical engineering, and he did power electronics, circuits that can convert electricity from one form to another. Those are the two professors I was impressed by. Of course, I spent a lot of time with Professor Rutledge. He's a high-frequency, millimeter-wave person. Electromagnetics. In my opinion, the top guy in electromagnetics and electromagnetic communication. I met other faculty from other departments not directly related to me. It was a busy day. I met probably 15 faculty on campus. I didn't meet Professor Mead because he was out of town, although I knew him. He wrote a book, VLSI, and I had a copy. [Laugh] But then, I wasn't really a circuit person anymore. Somehow, I didn't meet him. But I learned that Professor Rutledge continuously consulted Professor Mead about me.

ZIERLER: Did you give a formal presentation about your research?

TAI: Yeah.

ZIERLER: What did you talk about? What did you emphasize from your research?

TAI: I talked about micromechanics, a new field. Because when you talk about mechanics, people think of a pump. People actually think about big gears. People think about automobiles. People talk about transmission, frame. "No, no, no, it's micromechanics with a size comparable to transistors. And the future is really bright. We can talk about applications. There are so many applications already. I did talk about possible applications. But my PhD work was really just to create the possibility and demonstrated that we can fabricate very small mechanical parts that people know from cars. Except everything's down to micro, right? People actually think about robotic arm. Everything down to micro. People actually think about transmission, gears. But I make such little gears." That's what I presented, saying that, "People already spend trillions of dollars investing in semiconductor technology. There are all kinds of techniques. How to use chemical, how to use plasma, how to design a sequence so in the end, you have a transistor."

By the way, one transistor would take people several months to make. Except that, at the end of several months, you've got trillions of transistors together. What's the impact? The impact is, if that one transistor's useful, now I've got a trillion. Low price, mass production. That's what people want. That's how to push the world forward. Before semiconductors, Ford would build one car at a time. Boeing built airplanes one at a time. And then, of course, you'd have a production line. "This production line can build two cars a day." If you want to build 10 cars a day, what do you do? Five production lines. In a way, when you look at manufacturing, it's making one part at a time. But semiconductors are totally different. You make trillions of transistors at a time. If your customer only wants one transistor, you still end up with a trillion transistors. [Laugh] But that power is so big. That's why a chip can be so cheap. Same thing. how about micromechanics? If my micromechanical part's useful, it's exactly the same idea, mass production. We don't build one gear at a time.

Today, you build one gear at a time with a CNC machine. Two hours later, another piece of metal comes in. We don't do that. [Laugh] We work on the area on wafers. The potential impact is huge. And then, mechanics become micromechanics, and we know it can become nanomechanics. We foresee the world going into a realm where it's not just microelectronics, not just VLSI, not just circuits. Circuits side by side with sensors, with actuators. Actually, I call it combined microelectronics, microsensors, and micro actuators. I call it M-cubed devices. It's a M-cubed world. And then, of course, it'll become N-cubed. Nano cubed. Actually, one of my students stated a company called M-Cubed. [Laugh] It went IPO, but they changed the name. Anyway, that's the picture I tried to paint.

ZIERLER: Obviously, the Caltech faculty were excited about this.

TAI: I was lucky. For 10 or 20 years, I wondered, "How did I end up at Caltech? Everybody around me are all top guys. They're so successful. Why am I here?" [Laugh] Then, later on, I understood. The way I think, only new things attract me. I'm happy with new things. When someone says, "New," I feel happy. I can't stay with old things too long. It's not my personality. In a way, making micro gears, micro springs, it's something new. "Wow, I've never seen that before." I enjoy that a lot. And Caltech appreciates something at the beginning that shows the promise, something that will create a new direction that will never go away, and people will follow. That's my understanding of Caltech. Everybody should strive to do that. We're a small school. I didn't know. People told me Caltech was a small school, but I came, and I saw a lot of people. "What do you mean, small school?" Later on, when I learned more about Caltech compared to other big schools, I saw that we are very small. But that's how the small Caltech can compete with the rest of the world. I truly, truly understand. I fell in love with Caltech.

ZIERLER: To clarify, did you actually get offers from Bell Labs and Stanford?

TAI: Actually, I interviewed with Caltech. One week later, Professor Rutledge called me and said, "We want you to come again for a second interview. And you should bring your wife." He arranged to show my wife the housing. During the second interview, he told me, "We want to make you an offer." I actually talked to more faculty, and I only wanted to make sure of one thing. I said, "I need to build a clean room. I need to build a microfabrication lab because that's what I do." And I was promised that. Prof. Paul Jennings was the division chair. He told me verbally, eye-to-eye, "We will manage to build you a clean room if you want." I went back and talked to my Berkeley people, including my advisor. Bell Labs had already made me an offer. I wasn't that interested in traveling there. But I was interested in waiting for Stanford because of the whispers I'd heard that they were interested in me. I actually pushed a little bit. "How soon? I can't wait indefinitely." But I got word that Professor Paul Jennings said, "I will build you a clean room." Overnight, I made the decision to forget about Stanford. I called Professor Rutledge and said, "I'm coming. Your word's good enough. I'm coming." So I came.

ZIERLER: That's a perfect place to pick up for next time, when you start as a faculty member here.

[End of Recording]

ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Monday, March 20, 2023. I'm delighted to be back with Professor Yu-Chong Tai. YC, as always, thanks for having me in your office.

TAI: Thank you for spending time with me.

Joining the Faculty at Caltech

ZIERLER: Well, today, we're going to get to the main event, when you joined the faculty here at Caltech. Something we've talked about before, Caltech's smallness and the need for its faculty really to operate independently because there might not be 5 or 10 people for you to collaborate with here like there would be at a Stanford or an MIT. With that in mind, coming here as a new professor, what did you see as your area of specialty that you would carve out as an area of leadership for yourself?

TAI: I decided to join Caltech for two main things. One, Caltech treated me really well, Caltech made decisions really quickly. I sensed there was minimal bureaucracy, which I liked. The second thing is, I knew what I wanted to do. What I needed was a lab, and Caltech promised me a lab so I could build things. I knew what I could do. I didn't need to go to a big school. I didn't need that many collaborators around me. In a way, I believed this was a new field. People call it microelectromechanical systems. Basically, blending all the mechanical components in a micro world. There are unlimited possibilities. Coming to Caltech, in a way, was a blessing. I could focus on what I do, and nobody would bother me. And that's one really good thing that can happen to any faculty at Caltech.

ZIERLER: Not just in engineering, it could be anybody.

TAI: It could be anybody. As long as you know what you need to do, what you want to do, and if what you want to do doesn't depend on any other faculty, then Caltech will be heaven for you.

ZIERLER: In the way that Caltech really supports its junior faculty strongly, what was the process? Did you have a discussion with the division chair? Did you explain what you needed? Did the division chair just say, "Name a number"? How did you arrive at the resources necessary to build your lab?

TAI: Actually, in my case, I didn't negotiate that much. Professor Paul Jennings hired me, and he promised to build me a clean room lab so I could bring in equipment and start making small devices other than integrated circuits. And he made me an offer I thought was reasonable. I didn't really negotiate, I just took the offer, saying, "I like the way Caltech handles things, so here I am."

ZIERLER: What building did you start in? I assume Moore Lab was not here yet.

TAI: No, I started in the Steele Laboratory. That's where electrical sciences was. I was offered some pretty decent space in the basement. In fact, that's where I built my first clean room. Right now, that space is totally occupied by KNI.

ZIERLER: Tell me about building the lab. What instruments did you need? What was most important for you?

TAI: Although we're talking about making micromechanical components into the chip, in fact, the good thing is, I guess, nobody really needed top-of-the-line integrated circuit equipment, which are very, very expensive. Because it's also the beginning of a new field. There's fundamental science and engineering that needs to be done. In a way, the equipment I needed was really two or three generations behind the top-of-the-line equipment for integrated circuits. It was quite a story how I started, believe it or not. First of all, the startup fund from Caltech was never enough because I wanted to build a whole lab. But I did write a proposal to NSF. I remember very clearly my startup fund was around $400,000, and I wrote an equipment grant to the National Science Foundation because I was new faculty, and there was a special program. I wrote a proposal and sent it in, and I was granted another $400,000. A total of $800,000 was not enough. What I needed was $8 million dollars, not $800,000. It's never enough, but I decided to go for used equipment, which means maybe obsolete from the semiconductor industry. As I explained, all I needed was two- or three-generation-old equipment. I couldn't afford to buy new equipment. I started to chase after auction equipment. It's very interesting, the year I started, 1989 to maybe 1991, 1992, it was a bad time for semiconductors. Actually, the whole industry wasn't doing well, the employment rate was very bad. Actually, it was pretty awkward in Silicon Valley.

ZIERLER: What was going on then?

TAI: It was a bad economy. real estate went down. That was really a bad time. Actually, I said before that the time I graduated, 1989, it wasn't a good time. A lot of Silicon Valley companies were not hiring. That also meant a lot of small semiconductor companies went down. There were a lot of auctions. I remember very clearly. I'd usually go to the auctions to buy equipment. In a way, I was putting in a lot of risk because you'd never knew if it was going to work or not. But that was how I started. Some equipment, I paid only 10% of its new price. That was okay, we'd bring it back, take it apart, fix it, and put it back together, remove all the automated parts of it and put that equipment into basic manual mode. That's how I started in the Steele basement. Even with that $800,000, I was able to put together a lab and start my research. I had most of the equipment I wanted, and I started hiring graduate students. I actually even taught the first two or three graduate students exactly what to do in a clean room lab every day. That's how I started.

ZIERLER: When you say $8 million dollars, what would that mean? How expansive was your ambition in terms of what you wanted to do?

TAI: Actually, I never thought to have a big, grand, new lab. I only wanted to start a lab so I could start doing fundamental work. I knew there was so much fundamental work to be done. Even if I'd been given a big, grand, new lab, I wouldn't have been able to just jump to the development of a lot of useful devices. In fact, the whole field was going forward, and people were learning what was happening in the micromechanical world, how materials behave. If you make parts, how parts behave. This is very different from taking apart a car, looking at the gears, frame, or springs, and knowing exactly how they should perform. When you make something in such a small form factor, things actually don't behave in the ways we're used to. We have to learn a lot of things. I thought starting a small clean room lab was actually all I could handle. [Laugh] And that's really all I needed.

But I guess luck is always a part of my career. In 1996, I remember, Dr. Gordon Moore donated a big chunk of money for this building, the Moore Lab. I'd always been with electrical engineering, so I moved with electrical engineering from the Steele Lab to the Moore Lab. During that transition, there was another big opportunity. They built a bigger clean room lab for me, at least two and a half times the size of the clean room in the Steele basement. I was able to actually expand my equipment portfolio even more. But it's also very interesting, I had a bigger clean room, bigger group of graduate students and post-docs, and that was almost 10 years after I'd joined Caltech. I actually had a very clear idea of what this field could produce in terms of devices that could actually have some impact in the field. I expanded the capability of my lab and brought in more equipment. That was also the time I was able to get enough federal grants from various different agencies, NSF, NIH and even DARPA. The timing was perfect, everything fell together.

ZIERLER: The economy was better in the mid-90s also.

TAI: Oh, yes. That was a much better time. Much easier to get federal funding for research.

ZIERLER: Why is it so important for the lab to be a clean lab? What's the concern over contamination?

TAI: Although we're not making devices as small as transistors, we're talking about mechanical parts that are measured in microns and micrometers. That's one millionth of a meter. The size of a dust particle can range from sub-micrometer all the way to tens of micrometers. That means if you don't do the fabrication in clean enough environments, your device is either going to be covered by a lot of dust particles comparable to the size of the device, or during the fabrication, the device could never be realized because there are dust particles everywhere. A clean room actually is very important. Another reason I already knew at that time was, I'd already sensed that the biggest impact of this field, where we were exploring the micromechanical world, and later, the nanomechanical world, was going to include at least biomedicine.

It was very clear in my mind at that time. For example, there are very small pressure sensors that can measure your body pressure. In fact, people were already talking about a possible device to monitor a mother's womb pressure during delivery. There are challenges, for example, measuring brain pressure, the intracranial pressure. There are challenges measuring the interocular pressure, the pressure inside the eye. Because the pressure in our body usually relates to some diseases directly, if pressure regulation by our body goes wrong. That's just one example. There are many, many other devices I'm still dreaming about. Of course, devices for biomedicine usually take a much longer time. But at the same time, I thought it was the perfect research topic for a university. It's difficult, but it has practical impact, and it can end up in hospitals and human bodies to the extreme.

ZIERLER: To clarify, when you first started building your lab in the early 1990s, it was smaller, and you were more focused on fundamental research.

TAI: Correct.

ZIERLER: When the Gordon Moore Lab is built, and you have an opportunity to expand, now you're not only thinking about applications, but you're specifically thinking about biomedical applications.

TAI: Correct, that was around the starting time.

ZIERLER: If you were not thinking about biology generally, what was it about biology and bioengineering? Of all of the ways that you can have these micromechanical devices be applied or translated, why health sciences, why biology of all the applications?

TAI: Because I think the greatest impact for the integrated circuit was the transistor. Not only because they get smaller and smaller, but because they can be mass produced. When something can be mass produced and useful, you're going to change the world. Everybody can afford it. That's the biggest impact. And we see that from consumer electronics already. Every three to six months, there's a new product coming out. Everybody can understand that from cell phones. Every three to six months, there's a new model. Why? Because scaling makes the cell phone faster, makes transistors faster. Not only that, but because production can make it affordable to everybody. And you don't see that in other fields. But we see that very clearly from electronics, and especially consumer electronics. But you don't see that in most other fields. At that time, if you looked at biomedicine, it was evolutionary changes, very small changes. But biomedicine's actually the single biggest industry among all. It's very clear it can benefit from some technologies that can produce useful devices to replace a big, old thing or enable new capabilities that will be mass-producible and affordable to everyone.

ZIERLER: As your lab got larger, more applied, more focused on biomedical devices, did that change the kind of graduate student who was interested in working with you?

TAI: Oh, yeah, hugely. Especially at Caltech. There are always students who want to do new things. They want to do something that, when they talk about their work, nobody else has heard about it. [Laugh] It's Caltech's style. It's very different from Berkeley or Stanford. Over there, it's influenced by Silicon Valley a lot. A lot of students want to work with something and get hired even before they graduate. At Caltech, students don't talk about getting hired before graduating. Students want to talk about working on something within Caltech. I sensed that a long time ago, almost immediately after I joined Caltech. Caltech is the place pursuing new, creative, somehow crazy things, sometimes things that make people laugh. [Laugh] That's the way, I realized, it should be done.

Immediately, I jumped into biomedicine, although I'm an electrical engineer. I had to reeducate myself. On the other hand, I also realize I got a lot of good advice from people inside and outside Caltech. They said, "You should do something new or change direction every 10 years to expand yourself, open your eyes, try to tackle multidisciplinary, cross-disciplinary, or interdisciplinary research that involves different fields." And even at that time, places like NSF, the National Science Foundation, when they started a center, they wanted different people to work together. It's not single-discipline research anymore, they want multidisciplinary research. Everything falls together, as I said. I think about it, and today, it's quite amazing.

ZIERLER: Did achieving tenure free you up?

TAI: Not really. I heard the talk, "No matter what, get your tenure, then you're free to do anything you want." No, not so. In my own experience, before you get tenure, you work very hard, long hours every day. Absolutely true. Sometimes I tell people, "The worst job is not assistant professor, the worst job is assistant professor's wife." [Laugh] I thought, "After I get tenure, I can have more time to myself." No, after tenure, you become associate professor, and you're busier. When I was associate professor, I thought, "When I get full professor, life is going to be easier." Oh, no. [Laugh]

ZIERLER: You know what this means, you're going to go emeritus one day and be even busier. [Laugh]

TAI: It never stops. It amazes me how little time you can have. I always tell students, "I wish I had 48 hours in a day." But the only fair thing in the world, I think, is that everybody only has 24 hours in a day. Nothing else is fair. [Laugh] But it's fair that everybody has 24 hours a day. You just learn how to manage your time.

ZIERLER: In building the new lab, in the way you were saying the first lab, you only needed second- or third generation-old instrumentation, is the same true of the new lab? Or are you really in need of cutting-edge instrumentation at this point?

TAI: No, I still do not use cutting-edge instrumentation. Although, for research, it would be great to have. It's another new world. In the mechanical world, we usually deal with the large things, never small. If you go to a machine shop, every machine shop in the world is limited to around a millimeter. When you want to go smaller than a millimeter, we talk about the micromechanical world. That's where I focus most. Then, you go down, and you get to the nanomechanical world. In everything I do, I learned that if you couldn't master micro, it's very difficult to master nano, especially in mechanics. But then again, it should not stop people jumping directly to nano, starting to collect data, publishing what they saw, explaining things they're working on. That's what research is all about. But in my case, actually, playing with technology and being able to learn so much about materials and structure in mechanical form, I clearly see that there are a lot of crying needs from medicine. Especially after I started collaborating with doctors and biologists around me.

Those needs are obvious and should be met. It's time I really work with people to make working devices. Because one thing I realized is, if you read a lot of papers from the high-impact journals and all that, many are so exploratory. They could be promising, but who knows? It's a 20 or 30 years down the road kind of thing. On the other hand, I was facing a choice. I could continue to do that kind of research. But there were a lot of needs my collaborators wanted and didn't get, doctors wanted and didn't get, patients were crying for and didn't get. You could call it a middle-age crisis, but at that time, I decided, "Shoot, I'm going to try to apply my knowledge, and technologies, and smart graduate students to go in that direction." And there's no lack of interest from students and post-docs. They get excited when they hear, "We have an idea. There's a doctor we can work with. This can change medicine." Students come.

ZIERLER: When David Baltimore became president, famous biologist interested in raising the profile of life sciences here, was that relevant for you? Was that a good moment of timing for you?

TAI: It was one of the important factors. In fact, it started with President Chameau. And President Dave Baltimore actually focused a lot on biology because he's a biologist, a Nobel laureate. But President Chameau, and then President Rosenbaum, have continued to emphasize that societal impact is important for Caltech researchers. The timing was perfect for me, and I thought that was what I wanted to do. Then, I turned myself into a medical engineer.

ZIERLER: Entering a new field, when is it important for you to do the learning yourself, when are you hitting the textbooks like a student, and when are you collaborating with subject-area experts and can leave it to them? How do you draw that balance?

TAI: In my personal experience, I've found that, especially if you're talking about devices or technology, engineering is indispensable. I thought about how to reeducate myself, learn biology, all that. It turns out, in my experience, the best way to educate yourself is just collaborating with people. I collaborate with many doctors. I've learned so much from them. You may read 10 books, 100 papers, 1,000 papers, and not catch the important points. But when you collaborate with the right people, it doesn't take long. It takes just a few months for you to become an expert in the field. I think our education in engineering teaches every one of us, when you face a problem, you try to understand the problem first. I learned that a lot of doctors have a problem, but they may not spend enough effort to understand the sizes of the problem, the mechanics of the problem, the why.

"What are the equations that govern this? What controls the behavior?" Their lives are different. They interface with patients and all that. It's perfect. They hand over the problems, we study and understand the problems, then we communicate with the doctors. That is the best education ever. That's a shortcut. There's nothing better than that. The best way to "reeducate" yourself, to me, is to go directly to the doctor, collaborate with them, learn from them, ask them questions. I teach my students to do the same, and they feel exactly the same way. In a way, it's heaven at Caltech to be able to draw the best collaborators, to work with the best people. That's so true. [Laugh] There's not much else I can say. Work with the best people and learn from them.

ZIERLER: In entering this new field, I see there's one of two strategies. You can go broad and work on a lot of different biomedical engineering applications at once, or you can focus and master one area, then grow from there. What was your response? What choices did you have?

TAI: Both are great. There's no wrong in choosing either way. I think it's a personal choice. Some faculty do one thing through his or her whole life and become the expert. There's absolutely nothing wrong with it, especially if the field continues to open up new avenues. On the other hand, in my case, I think it's really my personality that dictates what I do. [Laugh] I love new things. I just like to tackle new problems. I challenge myself every day to think of something new that I haven't thought of. That's what makes me feel excited. I think personality is a very important part of it. Another thing that makes it easy for me to do that is, I'm more like a technologist. Technologists should always think about what you can build because technology's really there to build things. It doesn't make any sense to limit yourself to only one narrow field of things you can build. Technology is there to build everything. I like to explore new things. And I think the world runs that way, too. For example, when I'm thinking of extending the collaboration between two institutes, Caltech versus City of Hope, for example, or Caltech versus UC San Francisco, it's never one person's thing.

We started a symposium. Everybody agrees that for our potential collaborators together, they present their challenges, the problems they encounter, the puzzles they have. Then, we sit and try to figure out how to solve the problem, how to build things the doctors describe that they want. That way, actually, it's very difficult to limit yourself to a narrow field. I was stimulated, and I had really good collaborators from various different fields. As I explained, I learned that as long as you work with the best doctors, you learn so much. You don't need to worry that you don't know the new field. And I also learned that engineering students and post-docs can learn medicine quite quickly. Of course, they don't learn all the medicine, but they learn what they need very quickly. There's no fear. Actually, I think there's only pleasure and joy jumping into medical engineering. Everything falls together.

ZIERLER: You said as an assistant faculty member, you didn't need to collaborate much. In this new field, did you start working with professors in BBE?

TAI: Oh, yeah. I worked with Professor Richard Anderson. However, it's regrettable, but somehow, I collaborate a lot more with outside people, especially doctors. I think it's not by design, it's by natural evolution, because those are the people who continue to bring in real needs from hospitals. My colleagues in BBE are all doing amazing things. They also need technologies. There's no doubt about that. But they're actually applying knowledge and technologies to study biology more. On the other hand, there's a whole new world that people have unlimited medical problems that require good engineering to help. The way I see it is, medicine is really connected to how to treat or manage patients' diseases. Biology is usually trying to understand the fundamental biological sciences. Then, there are engineers. There's a triangle there.

If you connect engineering to biology, it's engineering for biology, which is what I do, bioengineering. But engineering with medicine directly linked together is what I define as medical engineering. It's very funny, there's also biology and medicine linked together. You need to study biology sometimes to create new medicine. That's how I explain different fields, who's in which domain. Here, definitely, I see huge opportunity for Caltech engineering to be connected to medicine directly, which has rarely happened before. It's a new opportunity. As I said, new things get me excited. I start talking to people, start a department, explain this philosophy, and most people agree with us. Here we go.

ZIERLER: What's an early example of when a clinician, somebody right there on the front lines, sees an opportunity or is frustrated by a particular device? First of all, how do they find you? How do you get word out that you're in this field?

TAI: Two things happened. Naturally, because we started this medical engineering, when people heard about that, they came to us.

ZIERLER: Because it's Caltech.

TAI: Why not try? Send an email. Because public media's so convenient, they don't even need to call. Most of the time, we don't answer the phone now. [Laugh] But email's so available, so powerful. I continue to receive a lot of email asking for collaboration. "I have this problem. How can you help me?" Including patients sending me emails. "Hey, I read this news. I'm looking for help." I bridge them to my collaborators, the doctors. That's one thing. The other thing is, through this mechanism that we arrange annual or biannual meetings with doctors. And because it's Caltech, I think, it draws a lot of attention. It helps so much, having the name Caltech.

ZIERLER: What's an early example of a doctor finding you and saying, "Here's this problem. Maybe you have a solution"? What was one early project that really set you on this path?

TAI: One example is my collaborator at USC, Dr. Mark Humayun. In fact, he was the first person I collaborated with. Because of meetings, conferences, we'd get together, and we had a lot of time to brainstorm together. He's an eye doctor, ophthalmologist. He probably has hundreds of different needs. [Laugh] He's an MD/PhD, so he knows engineering. That's one example. The moment we started talking, there were so many ideas. We actually talked about devices or some interesting ways to treat eyes, including cataracts, glaucoma, macular disease, even diabetic retinopathy. All the serious diseases. They all actually can benefit so much from medical devices. In fact, it's a pretty big industry. If you talk about eye devices, people already know contact lenses and lens-based devices. There are so many. It's a big industry. I've been working with him for 20 years simply because the first time we met, both of us realized it was a perfect match. I build small devices, and he needs small devices to treat eye diseases. We've coauthored maybe hundreds of papers together. Even today, we work very, very closely together. Maybe five companies spun off from Caltech where he's a cofounder. It's not just Caltech, they're Caltech-USC spinoffs.

ZIERLER: What would be an example of increased technological capabilities in integrated circuits or electrical engineering that were most useful as you jumped into biomedical engineering?

TAI: It's super useful. Maybe 40 years ago, the term smart sensor came out. Smart was very clear, it had to have the capability to do signal computing, information management, all that. Today, you can call it AI sensor. Everybody talks about AI. Actually, no one understands what's an AI sensor. [Laugh] But even 30, 40 years ago, we were talking about smart sensors, and then we saw it. It means electronics should be together with a sensor, which can be chemical, mechanical, or electrical. And mechanical actually became very useful. There are tons of examples. Just in our cell phone's microphone, it's all integrated. Electronics, transistors working side-by-side with a membrane, which is a mechanical gadget. It's really just a membrane. The sound wave comes in, the membrane moves, and electronics pick up the signal. That's today's microphone. It's all integrated, what I describe as M-cubed. There's microelectronics and transistors, there are microsensors that sense things, and there's also micro actuators. Because things move, you have to use actuators. Another example is the accelerometer. Same thing, it's all integrated accelerometers. Electronics have to be next to it. In fact, people in electrical engineering understand that if you don't have transistors exactly next to the sensor, you lose a lot of sensitivity.

Same thing with gyros. A magnetometer can sense a magnetic field. In the traditional sensor field, people talk about weather stations. A micro weather station can sense humidity, pressure, altitude. Gas sensors are sensitive to CO2, maybe nitrous oxide. There are unlimited things. They're all smart sensors in the sense that electronic computation must be next to it. Today, we call it AI because AI seems to do more than just transistors. It can do machine learning, it can do deep learning. Just data-based, not physics-based sensing. I'm doing that, too. It seems to make sense. Believe it or not, it's a blessing that we see that happen. It's a no-brainer, people agree. For the last 30 years of my career, I've seen it exploding. There are more. In fact, the same thing's going to happen to biomedical devices, I guarantee you. Except that in biomedicine, usually new technology is always 20 or 30 years late. Part of the reason's the US FDA. The FDA rules make change of technology very difficult, unlike the consumer electronics industry. Electrical engineering plus medicine, oh my God, it's unstoppable. How can you stop that? It has to happen. [Laugh]

ZIERLER: When the Kavli Foundation folks started talking to Caltech about building KNI, was that relevant for you? Did that present you with new opportunities?

TAI: Yes, and no. Yes, in the sense of going into nano because they focus on nano. That was the height of the nanotechnology wave. I actually do nanotechnology. I play with new nano materials. Of course, it's promising for biomedicine, too. There's no doubt in my mind it will continue to take many years of research. As I said, in order to make something into real life, you really have to master it, not just understand it. You have to also know how to manipulate it, how to fabricate things, especially if you're talking about materials. And nanotechnology has the biggest impact in terms of producing new materials. At that time, actually when I focused on real medical problems, I realized nanotechnology is still far away from impacting real medicine, although it has great potential and promise. Microtechnology is what people need today.

ZIERLER: Nanotechnology is not developed yet? Is that what you're saying?

TAI: People don't quite understand it yet. That's the thing. In medicine, it's harder. You don't just try it because there's regulation, safety concerns that slow it down. It will take time for people to master it and be able to answer a lot of question. Of course, the first question the FDA always asks is, "Is it safe?" Today, you can ask any nanotechnology guy, "Is your stuff safe?" [Laugh] It's hard to answer. That's a reality we just have to face.

ZIERLER: Did your work with sensors make JPL an asset for you given all of the sensor work that they do?

TAI: I know people who are interested in biomedical devices from JPL, but JPL is an institute that has a very clear mission. It's for space. Unfortunately, it's hard for my friends at JPL to only focus on biomedicine, not somehow do their mission. It's still true today. In the end, I think there was less and less connection between Caltech medical engineering and JPL, which is unfortunate. Their mission is very, very clear. Their managers are mission-focused people.

Formalizing Medical Device Research into a Department

ZIERLER: When did you have the first idea that maybe Caltech should not just have you in your lab, but a department or center of medical engineering? How did you develop that idea?

TAI: It was 2011/2012. Clearly, there were many faculty who thought their research could expand into medicine. Like Professor Axel Scherer. He's very active in building devices for medicine. Professor Mory Gharib, Joel Burdick. Professor Emami. Professor Hajimiri. In fact, the more faculty we talk to within engineering, the more faculty agree that going into medicine is a great opportunity.

ZIERLER: There was already a critical mass of faculty research here.

TAI: Yes. Actually, in 2012, we decided to coauthor a white paper. There were 10 faculty, including all my colleagues I mentioned. We wrote a white paper to the engineering division, EAS. Professor Ares Rosakis was the divisional chair. It took him about a year for him to decide, "I'll back you up. Go do it. Whatever's needed to be done, I'll back you up." We started the effort to start a department, including with a preparation to present this idea to the faculty board. That was in 2013.

ZIERLER: Over that year, did Ares really need to be convinced?

TAI: Yeah, it took him a year. It's quite an ambitious proposal, to start a new department.

ZIERLER: Going back to our first conversation, there's the concern that the Harvards and Berkeleys have been doing this for a hundred years.

TAI: That question needed to be answered. We had answers, but it still took people some time to digest the idea.

ZIERLER: The fact that Caltech does not have a hospital, I understand you see that as an asset, a good thing for you.

TAI: I see that as a good thing.

ZIERLER: But for the people you had to convince, division chair, provost, president, was that a source of concern? And if so, how do you convince them it's not?

TAI: Oh, yeah. I talked about what happened in the faculty board meeting. I presented the idea, and there were many questions. That was one question, that Caltech didn't have a medical school. I told them, "Not having a medical school is actually an advantage. That means people will not worry too much. In fact, people will feel really good just knocking on our door. They wouldn't have competition." I actually sensed there was a psychological worry. If you work with another institute, and the other institute has other people who are sort of your competitors, you always worry that your idea can be stolen away. [Laugh] You can lose it to the institute because they can form the collaboration right there. Why do you need you? We don't have a medical school, which, it turns out, makes people feel free, comfortable. It's not likely that we can steal their ideas. It's harder. Also, we sort of manage that. When we talk about collaboration, when we arrange a symposium and all that, we sign NDAs to make everybody feel comfortable.

When we have the biannual meeting with UC San Francisco, every attendee signs an NDA. We establish this firewall to make everybody feel comfortable speaking freely. Of course, Caltech has always run on the honor code. And they trust us even more. It's never happened where an idea was stolen, and people regret having talked to Caltech. We did everything carefully. In a way, my colleagues at Caltech understood this sensitive psychology. They actually agree. The most important, as I described before, I said to my colleagues, "Turn that question around. If you are a medical doctor, you have a problem, and you have some ideas, you need to work with engineers, do you want to work with first-class or second-class engineers?" It's not a question, really. [Laugh] I told them, "We're the first-class engineers," which they couldn't deny. I said, "As long as we let people know we're starting this department, and we're open for collaboration, what do you think will happen? People are going to start knocking on our doors." That's exactly what happened. The most frequently heard complaint from our faculty is, "Our hands are so full, we can't take more projects. We have too many people who want to collaborate with us."

ZIERLER: That's a good problem to have. Was a good relationship with the Office of Tech Transfer important for setting up the new department?

TAI: It's hard not to have a good relationship with them. In fact, I remember close to 20 years ago, Larry Gilbert was hired from Boston University. Immediately, I heard the news that, "Wow, they want to change OTT completely." I'm a close friend of both Larry Gilbert and Rich Wolf. At that time, the two of them ran the office. It's hard not to be close to them. As soon as you have really good stuff in biomedicine, you're going to be connected to them. And they're so friendly. They've kept their word that they would help us, help faculty, achieve whatever we want, especially in technology transfer. They really did. They're so friendly, the situation totally changed. You walk in there, and you feel that you're welcome.

ZIERLER: I wonder if you can compare the experience. You're executive officer for electrical engineering, a very old discipline at Caltech, from 2005 to 2008. Then, you're executive officer for medical engineering, a brand-new discipline, from 2013 until last year. How do you compare those two different responsibilities and roles?

TAI: When I was executive officer of electrical engineering, it was a big department but on the right track already. Being executive officer, you don't need to make that many new decisions.

ZIERLER: Just don't break anything.

TAI: Yeah, just do as it is. Secretaries of electrical engineering ran the department, really. [Laugh] They can walk on thin ice. They're more experienced than the faculty. I only needed to spend a few hours a week on average, not that much time, just making sure I knew what was going on. There was not much decision-making. It was obvious. In fact, we had a really good building manager, Carol, and Linda was the EE department secretary. The two of them were the executive officers, in my opinion. They ran the department. When I started medical engineering, it was day-and-night different. I had to make every decision. And not only making decisions. When something comes up, for example, that we don't have enough courses, it's not a decision of whether we want to keep or stop a course. The problem's always starting from zero. It's to make zero to one. Or maybe zero to 0.5. The headaches are totally different. It's too crazy. Including actually finding a secretary, which turned out to be my personal secretary, Christine. Quite an experience, starting from nothing, to make a department.

ZIERLER: Did that include faculty hires?

TAI: Yes.

ZIERLER: How many new lines were you able to get?

TAI: For the last 10 years, we were able to hire Professor Lihong Wang. That was quite a story. We started the department in 2013, and I was told that there was a senior faculty position on campus. And very rarely does Caltech target senior faculty. Usually, Caltech's priority is to find the smartest young people. I was told by my division chair, Dr. Rosakis, "Hey, there's a campus-wide faculty search for senior faculty." We got the news late. The other divisions heard the news earlier. They all tried, I believe. But the position was still open. My division chair asked me to do something. We looked around, and it was suggested we should look at Professor Lihong Wang at Washington University in St. Louis. He was with biomedical engineering there, already a famous guy. In fact, many years before that, he even visited my lab at Caltech. Anyway, he realized the first photoacoustic tomography machine. I put together a package. He's really strong. Not just his research, he got letters from Nobel laureates.

We succeeded in stealing that position, so we hired Professor Lihong Wang. This was 2016. It was an opportunity that fell from the sky. Usually, it's BBE, or CCE, or PMA, physics, chemistry, or biology that gets a position like that. But that was a big success for engineering. I was so happy. And Professor Lihong Wang is the pioneer of photoacoustics. [Laugh] after him, we hired Professor Wei Gao. We call him the sweat guy. He's the sweat analysis guy in the nation. There's absolutely no regret that we started medical engineering. We believe it was the right thing for Caltech. During the faculty board meeting, my colleagues voted for it. Unanimously, they voted yes, not a single abstain or no vote. It was the right thing to do because we're going to do it differently. I tell my colleagues and students, "Everybody else's engineers are trying to learn medicine. That's the wrong way. We're engineers. We do engineering the best. We should continue to do what we do best, which is engineering. We want solid engineers to be our faculty."

I didn't want to hire a medical doctor to be our faculty. That's not our strength. And we continue to recruit engineering students. We emphasize that their education should be engineering. You look at BME's students these days, they're a little bit medicine, a little bit biology, a little bit engineering, all at shallow levels. I said, "No, we want to train our students to be very solid and deep in engineering." Because I really believe all medical problems are complicated and multidisciplinary. You have to go down very deep to understand and solve the problems. I want us to be different, but different in the right way. We had the chance to do that. I tell my colleagues, "Please understand not only that our research is different. In fact, the way we train students and post-docs, the way we solve problems are also different." That, to me, is more important. Because in the long run, people will understand. "Whoa, what happened at Caltech? It's a young department, but they continue to prosper and solve the problems the other schools couldn't do."

ZIERLER: Were you involved when the Chens started becoming interested in supporting Caltech?

TAI: Yes. I heard a story, but the way I'll put it is, engineering has very little to do with that interesting case. It was handled, and nicely, by mostly BBE faculty.

ZIERLER: Did the Chens specifically want to support medical engineering?

TAI: No, they wanted to support neuroscience. The story is that Mr. Chen himself is crazy about the human mind, the brain.

ZIERLER: You're talking about Chrissy and Tianqiao Chen.

TAI: Yeah. Caltech received a solicitation at some point. He sent out an invitation to many other schools and institutes, including Stanford, MIT, all that. Caltech actually answered and sent in a proposal, but with an attachment saying, "If we can get more money, we can do this and that, greater things." That got Mr. Chen's attention. Originally, it was a much smaller proposal, something like $20 million, which then expanded into a $100-million proposal. But he, himself, wanted to do neuroscience. There are only a few faculty over on the engineering side that collaborate with the neuroscientists over there. Medical engineering leans more toward medicine, much less in biology. Personally, I have a lot of interesting neuroscience, and part of it is through collaboration with UCLA and USC, but more on the immediate treatment side. For example, I build devices that can do neuromodulation, neurostimulation, implanted in the eye with the goal of bringing back loss of vision for blind people. There's some connection. But neuroscience and neuroengineering hasn't been on the radar for engineering for a long, long time.

ZIERLER: Is that because the brain is still so poorly understood?

TAI: No, actually, a lot of neuroengineering can be done. It's just that Caltech actually never hired engineering faculty directly from neuroengineering. I like to talk to people and think about what's possible, so I work with neuroscientists. Last year, we tried to hire a female faculty from Harvard, and we made an offer. She's in neuromuscular sciences. But she also got an offer from Harvard, so she stayed there. Absolutely, within engineering, I think the intention is to expand and bring in new fields. But it's a slow process.

Partnership with the Cherngs

ZIERLER: Were you involved in bringing the Cherngs to Caltech?

TAI: Yes. I have known Andrew and Peggy Cherng for a long, long time, even before we started medical engineering.

ZIERLER: What was the connection?

TAI: We all belong to the Asian community here. Once in a while, there would be some occasions where we'd meet each other. Or a gathering at friends' houses or what have you. We knew each other. We weren't close, but we knew each other. After we started medical engineering, the person who helped the most is the prior chairman of the board, Dr. Dave Lee. He's known the Cherng family very well long before we started working with them. Dave Lee and I have been neighbors since we moved to Caltech. In fact, my first house, a very small house, was bought by Dave Lee's wife, Helen. She was a realtor then. She actually arranged to take my wife to see housing during the second interview I had with Caltech. Dave Lee and I are close. And the house Helen found for us within the same block of Dave Lee's house. [Laugh] His dog always ran out and came to my house. [Laugh] Anyway, the connection definitely was there, and that's because of Dave Lee. Then, we started medical engineering, and everybody automatically thought the Cherng family should be a target. Dave Lee worked with President Rosenbaum closely. In the end, though, it was Dave Lee that made it happen. It was all to his credit.

ZIERLER: Did Peggy and Andrew have a specific interest in medical engineering or just wanted to support Caltech in some way?

TAI: They actually specifically asked for medical engineering.

ZIERLER: Do you know what their interest is specifically in that?

TAI: Part of it is Dave Lee. Part of it was that we'd just started the new medical engineering department. It was a very good opportunity for them to be very helpful to medical engineering. Within our community, I think Andrew and Peggy both experienced friends and family having diseases. In fact, including City of Hope, there was a small Asian community that would handle medical problems of friends and family. I think that's really the main reason. Believe it or not, the Chinese community around Caltech are heavily into medicine. A lot of medical doctors. I don't know why. But most parents don't mind their kids getting into medicine or becoming doctors. Medicine actually is something that got their interest. They were more exposed to it.

ZIERLER: What did Peggy and Andrew Cherng's support make possible?

TAI: The real money started to come in 2017, and my decision, together with the division, was to spend all the proceeds on students and post-docs as much as possible. My hope is really on the young students.

ZIERLER: This attracts the best students and makes their lives here easier.

TAI: Yes, exactly. They're the ones who do all the work. [Laugh] We're able to offer graduate students first- and even second-year tuition. We're able to offer the smartest students an invitation to Caltech without them worrying about money problems right away. We tell these young students, "If you come, you'll have at least two to three years of time to develop what you want to do." In most other schools, students get locked into some sort of funded project, which means they don't get to choose what they want to do. They're assigned to projects they have to do. Except for those very rare cases where a student gets a national scholarship or fellowship, where there are no strings attached. The percentage of students that happens for is very small. But because of the Cherngs' endowment, which we decided to spend mostly on graduate students, we're creating opportunities for the smartest students to come to Caltech and work out the research they want to do. We don't lock them into something even before they come.

That's the best way I can explain it. Allow students to develop their minds, talk to doctors, and figure out what problems they want to work on, what's a match for them. That's it. That makes our program so unique in the whole world. In fact, when endowment return was good, we were even able to fund some graduate students for three years and give them the best program they wanted to do. And it shows. Many of our students start their own companies or run spinoffs from the department. You can tell, it glows in their eyes, they're doing something they want to do. I'm telling you, I come from Berkeley, and when I was a student, there were so many students fighting for so few positions. [Laugh] You were lucky if faculty would take you in and give you something to do. When I was a graduate student, I was poor, couldn't afford to pay tuition. The first opportunity, I was offered a GRA, which paid my tuition, and I'd get some stipend. I took it without any hesitation. [Laugh] I think what we do is really the right thing to support our students, give them the chance, train them on how to come up with a good research proposal by themselves, then actually interface with doctors and clinicians, with the hope that they can really solve problems.

ZIERLER: In the last 10 years, undergraduate and graduate students have really become interested in computer science at Caltech. Have those interests filtered over to medical engineering? Has medical engineering become more computational because of the kinds of things and connections students are seeing?

TAI: Not yet, but it's happening. For example, with graduate student applicants, there's been a shift, although it's still only 10 to 20%, that students have learned AI and computer science and want to apply it to medicine. This year, I've seen three or four students who have that in their background, and that's what they want to do. But so far, all faculty in medical engineering, including the new hires, Professor Wang and Professor Gao, as well as myself, are all physical science engineers. We don't touch much of the artificial intelligence. But things will change as there's more and more demand from the students. That's the most important factor that will push this forward. As more and more smart students want to get into that, some faculty will start doing it. And this is backed by our current divisional chair, Dr. Atwater. He's one of the biggest supporters of AI. He's pushing for AI programs, trying to bring AI into engineering. And I see that happening. I myself am learning AI. But I don't want to just jump into AI and start thinking where I can use AI, but I want to form teams that have AI factored into them. Find a real problem with doctors, maybe start a project that AI can be used for. I think that's actually the best way to do it for now.

ZIERLER: Moving our conversation closer to the present, when COVID hit, what did that mean for the department and your lab?

TAI: Well, it slowed down things, for sure. [Laugh] Actually, it didn't do much harm to me personally. I know it affected other faculty. The larger research groups got hit the hardest. There's one thing I actually didn't like. For a long period of time, 10 years, I've had to take care of the medical engineering department, and I've had less and less time to do my own research. In a way, it's been my decision. I really want to make medical engineering good. I promised myself I would put the priorities of the department above my research. Otherwise, who knows where it would go. Over the years, my research group's shrank a bit. It's always a good size, though. But I wasn't affected by COVID that much because during COVID time, Caltech allowed students and post-docs to work in the lab as long as the proper "spacing" was observed. Because my group's not that big, social distancing can easily be followed, so it didn't hurt me that much. I didn't have that many students, and I had enough lab to provide safe spacing for all the students. But you can imagine a faculty with a big group, you couldn't get the right spacing. Now, I'm ramping up my group, but the pandemic's over. Hopefully. [Laugh]

ZIERLER: Did COVID as a medical crisis create new research directions for you?

TAI: It's interesting, it's linked to one direction I've been doing, and I've collaborated with Dr. Yuman Fong at City of Hope. COVID came up, and people paid a lot of attention to viruses. It turns out, one field that got more exposure is using viruses to treat cancer. In fact, there were a couple reports, I know one is from the United Kingdom, that a cancer patient mysteriously got COVID and found out his cancer was cured. This is a real story, you can Google it. In fact, to my collaborator, that's not a surprise because people have been using viruses to treat cancer in research for a long time. Amgen had the first virus drug approved by the FDA to treat melanoma, skin cancer, the worst kind. It's a field called oncolytics. Literally means using tens of trillions of viruses shooting through your bloodstream.

Genetically engineered viruses, though, not COVID. They seek and kill only cancer cells, leaving healthy cells untouched. It's very interesting. I had a student from the first class of medical engineer students, who had a background in tissue engineering. His name was Colin Cook, a smart lad. Because his background was tissue engineering, he'd learned to grow cells and all that. We invented a new way to improve cell growth, cell culture, then later on, virus production that's almost 100 times better than the current technology. Obviously, we worked with OTT and spun off the company, and Colin is pretty much running it. One big application we're focusing on is using viruses to treat cancer.

ZIERLER: Good luck.

TAI: Yeah, we need it. But we believe in it. We want to be the best cell- and virus-production company. To produce viruses to treat cancer is our number-one target.

Always Target the Biggest Challenges

ZIERLER: Moving our conversation right up to the present, you stepped down as executive officer in 2022. You already mentioned your research is ramping back up. What are you now able to focus on? In what ways is your research ramping up?

TAI: I was able to recruit students to always target new problems or the most difficult unsolved problems. I actually have been feeling happy for quite some time now because most of the time, I can focus on research, thinking of new ideas. I'll give you some examples. My most senior student, Allen Shang, actually is dealing with diabetics. There are two types, type 1 and type 2. Type 1 is autoimmune, kills their own beta cells. Type 2 is insulin resistance. Both types need your body to produce insulin. One field that's grown really well is to use stem cells to differentiate into beta cells. People have tried this for a long time. There are two things they've tried. One is donors' pancreatic islets, basically looking at the beta cells. If you have a donor, you can put their beta cells into your body. Or you can put stem-cell-derived beta cells into your body. They see a lot of immune response. When they do that, the number-one problem we've discovered is how to keep them alive for at least the first few weeks due to lack of oxygen. It doesn't matter what tissue is put inside your body, the first thing the implant faces is ischemia, not enough oxygen, and it dies.

This is related to any tissue implant, unless it's a working implant, because the first thing you do then is connect a blood vessel. But we're not talking about the whole organ, we're talking about cells. How do you bring oxygen and nutrients to them? That's the number-one problem. The second thing is keeping them away from the immune system. There are lymphocytes, white blood cells that come to attack. It's a collaboration with City of Hope. We're making micro implants that can actually keep the implanted cells alive and happy, and keep them away from immune problems. We've actually had quite good success at showing that if we bring oxygen to them right away, they'll survive. That's one project. I have a second student, Suhash, and we've brainstormed for two years. We ended up collaborating with Professor Berger from UC San Francisco. He's been treating brain tumors, GBM, one of the worst kinds. We have a pretty good idea to try and see if it works, to attract the cancer cells to a device and kill them. With a third student, again after two years, we've decided to explore medical acoustics. Our body produces all kinds of sounds. It's inspired by my mechanic. They fix cars because they listen to the sound.

ZIERLER: We can do the same for the human body.

TAI: Except that we need a tool. Our body produces a lot of sounds, but usually, they're very weak. We're developing the most ever sensitive acoustic sensors, specifically targeted for medical acoustics. We're trying to make it 100 times more sensitive than today's microphones.

ZIERLER: For the last part, to wrap up this excellent series of discussions, I want to ask a few retrospective questions about your career, then we'll end looking to the future. What lessons have you learned in the various ways your career trajectory has unfolded? Coming to the United States, getting involved in electrical engineering, seeing the value of Caltech, entering into a new discipline. When you step back and look at all of it, what are some lessons you've learned from a life in engineering and academia?

TAI: To make the answer short, I'll answer this way. I've learned, and this is also my personality, to try to do something different and be bold. Never chase after a hot field. A lot of friends of mine have tried to get into something hot. I've learned since I was young–during my graduate study, I started as a semiconductor engineer, learning transistors. I told you the story that triggered my mind by an officer from the Army's Henry Diamond Lab asking about making detonators. Suddenly, micromechanics came in. "Oh my God, that's it. Something new with a lot of possibility." We tried something, and it worked right from the beginning. That was what I decided to jump into because it was something new. Never been done before. Then, I joined Caltech, and of course, the number-one problem was to bring in students and get research money.

Very quickly, I realized that the United States is wonderful. It's a blessing that there are many funding agencies you can write proposals to. I've found it's more fun to write proposals on a subject that would surprise most of the reviewers. Actually, that strategy may not work for NIH. The NIH is more conservative. But it works perfectly for DARPA or sometimes the National Science Foundation. There's always a chance for something new you read and think, "Oh my God, I've never thought about this." All my life with Caltech, I've been pursuing that kind of research. Very new, something that's never been done. Of course, later on, I found out that most faculty at Caltech feel the same way. I was lucky that I've never run out of ideas because I keep looking for new directions. There's always a new challenge. Do new things. Don't do the same thing you did yesterday.

ZIERLER: Let's say you did not come to Caltech, you went somewhere else. Do you think the pull of biology and medical engineering meant that you would've entered this field no matter where you were, or is there something really unique about Caltech that pushed you into this field that wouldn't have happened otherwise?

TAI: I definitely believe that I wouldn't have had the kind of freedom I've had at Caltech elsewhere. For example, if I'd joined Berkeley or Stanford, they're big schools, lots of people, that even hire similar people, similar background. For example, MEMS. When micromechanics were hot, all these schools usually hire three to five faculty. Immediately, they become your competitors. At Caltech, every field, we hire one person. There's no one who competes with me at Caltech. The direct benefit is, I get all the attention and resources that could come in to micromechanics. And obviously, if I was with Berkeley or Stanford, my career path would be totally different. I wouldn't be able to do all the crazy things I could do at Caltech.

ZIERLER: You have students who have gone on to very important academic careers, and you have students who have gone on to very important careers in industry. I wonder how you compare and contrast the pride that you feel in those two different kinds of students and the career paths they've taken.

TAI: Honestly, I don't think too much about what they choose to do. I think it's totally their choice. I think my responsibility and pride comes from the fact that all my students are trained really well, so whoever deals with them will say, "Whoa, these are good engineers." That's what I care most about. If you want to go into detail, the most important part is that they learn to look at problems deeply, to fully understand the fundamental part of it, so they can solve problems systematically. In many cases, they can solve a problem this way only. Because if you're not trained to do that, you wouldn't be able to solve some very difficult problems. I actually care deeply for all my students to learn the capability, to be a solid engineer in the sense that they can solve problems others can't.

ZIERLER: And this is true both for your students who have gone on to academia and industry.

TAI: Correct. No doubt about that. In fact, if I think back, I have more than 10 students in academia right now. They're all doing really well. [Laugh] However, I had a little bit of concern for them that not all of them are able to look for new directions. For those who couldn't continue to look for new directions, I think they all suffered, to some degree, at some point. Some people are born to do that all the time, break boundaries, gather bruises and cuts, no regrets. Some people would rather stay comfortable.

ZIERLER: I know for you, it's very nice when the research does get into applications, but your primary motivation is in the fundamentals. With that in mind, what has been most intellectually stimulating for you to learn about how these systems work?

TAI: There's no one thing I can point out. But one most important thing is, whatever you don't know, face it and learn it. There's no fear. Don't tell yourself it's not your field, you don't need to know. There's never-ending learning. It's so true in research. You can't expect that the next research is always within your comfort zone.

ZIERLER: And if it is, why bother doing it?

TAI: Exactly. In fact, I've been telling students, "If you feel your life is hard or you're puzzled all the time, that means you're learning. If you're not puzzled any day, that means you're not learning, not improving." In my case, actually, there were days in the early times it was painful, but later on, you learn that's the most enjoyable thing you do. In fact, right before you came in, I was studying the human brain's reward system. I may get into something like that, using deep-brain stimulation with electrodes. And I learned that where to stimulate is the most important thing. In theory, electrical stimulation can help all brain diseases that data can't do anything about today, including Alzheimer's, depression, anxiety, drug addiction. I had a lot of fun before you came in studying the reward system. [Laugh] So it's continuous learning.

ZIERLER: With that in mind, last question, looking to the future. Because you always want to do something new, what's the frontier for you? What's the next big project for however long you define the next chapter of your career?

TAI: Believe it or not, there are two things I'll continue to do no matter what. The first is implants, devices that can be put inside your body. Everybody watches sci-fi movies. It turns out, for medical devices, that's the last frontier. The worst place to go is inside our body, putting devices inside the body. But then, it fits everything I do. First of all, you want everything to be small. You don't want to put a baseball inside your body. Golf ball, maybe. We actually want devices the size of a grain of rice. Everybody's going that way. But when you think about implants, there are a lot of fundamental sciences you have to learn, otherwise it will never work.

Our body hates anything being put inside it. That involves a lot of biology, the immune system and all that. To make good, interesting, small biomedical devices inside to body will continue to be on my radar screen. I want to continue to do it. I've been doing that for a couple decades, and I've learned a lot through a lot of failures. [Laugh] And through a lot of success, too. But I think I've reached a level where there are many things we can try that have not been tried before. That's one direction. The other direction I'm fascinated in is treating brain diseases. I will go into that, trying to treat brain diseases. I think I can actually do a good job in terms of making use of implant devices for neural recording or neurostimulation.

ZIERLER: Devices that go right into the brain.

TAI: Yes. I really believe I have something that's worth trying, that's very new, no one has done it before, in terms of how to protect devices that have all the computing powers and electronics. It turns out, that's the most difficult thing, when you talk about implants. If you want to put a battery in the body, the FDA is not going to be easy about that. Then, you ask questions. Electronics, transistors, computing power, AI. All those need to go in. There are a lot of implants already today that were pretty much developed in the 60s by Medtronic. It has not changed a bit. It's time to change it.

ZIERLER: This has been a phenomenal series of discussions. I want to thank you so much.

TAI: Thank you, David. It gave me a chance to speak from my heart. [Laugh]

[END]