She Lights the Way
Transcript for Interview with Penporn Koanantakool
[00:00:00] Nicole Huesman: Penporn, hello, it is so wonderful to have you here with us today.
[00:00:05] Penporn Koanantakool: Hi, Nicole. It's so nice to be here as well. Thank you for having me.
[00:00:09] Nicole Huesman: So can you tell everyone a little bit about what you do?
[00:00:14] Penporn Koanantakool: Oh, yes. So I'm a software engineer at Google. So I work on AI software infrastructure, which is the basic building blocks to run AI and machine learning programs. So I focus it on making it run faster to save time or to be able to do more within the same amount of time. Oh, and I think I should also say here that what I talk about in this podcast is my personal opinions and it's not related to my employer.
[00:00:41] Nicole Huesman: Excellent. And thank you for that disclaimer. Penporn, can you tell us a little bit about your background?
[00:00:46] Penporn Koanantakool: Sure. I was born and grew up in Thailand. I did my undergrad at Kasetsart University, also in Thailand. I was majoring in computer engineering, and then I came to the U.S. for my [00:01:00] PhD in computer science at UC Berkeley. I was specializing in high performance computing, which is using many computers to make programs run faster in parallel. And after graduating, I joined Google to work on this AI infrastructure.
[00:01:18] Nicole Huesman: Really cool. So, You moved from Thailand to the U. S. Can you talk a little bit about what your motivation was for that move? And then, what was it like to adjust to things in the U.S.?
[00:01:35] Penporn Koanantakool: So yeah, in middle school, I followed my cousin to travel in the U.S. And that was the first time I got to know the U.S. when I thought it was a nice country. And then later on, I realized I wanted to pursue higher-level education in a U.S. university as well.
[00:01:52] My language was a pretty big barrier. So in my first one or two years, I [00:02:00] couldn't recognize a lot of what people say and it took me a really long time to form sentences that I wanted to say as well. So it kind of reduced my opportunity to interact with other people, like, say, when we are having lunch with other people and they are having conversation about some things. Sometimes I don't even catch what they're talking about, so that's one thing. But when I catch it, I have to take some time to think about what I want to say, and then try to translate it into English. And by the time I'm ready to say it, the conversation has already moved to something else. So yeah, it was pretty tough in the first few years.
Nicole Huesman: How did you overcome that?
[00:02:40] Penporn Koanantakool: I have really nice and friendly people around me who will try to looped me in many things and and the more I got interacted with them, the more I have the courage to speak and one of my friends asked me to help coordinating this kind of networking lunch as well. So, women in the [00:03:00] EECS department electrical engineering and computer science department, we have weekly lunches together. And my friend was the original coordinator of the lunch. She asked me to help take it over. So I did that for about a year and it got me into talking with more people in, in, in the group and yeah, and being more social. So I think that helped.
[00:03:19] Nicole Huesman: That's wonderful that you were pulled into that group and that, you know, it made you feel more comfortable. It sounds like it also provided you with confidence.
[00:03:29] Penporn Koanantakool: I think it made me less tough on myself. Yeah. Knowing that it's actually okay. Who cares? Yeah. You just talk and say what you need to say. Yeah.
[00:03:38] Nicole Huesman: One of the things that really spoke to me and it just seemed like a beautiful story is what your path was into coding. Can you talk a little bit about that?
[00:03:52] Penporn Koanantakool: Yeah, so, one of my high school senior friends, he participated in a coding competition. It's called. It has a [00:04:00] nickname of Computer Olympics, but I think its official name is International Olympiads of Informatics. It's just like a computer programming competition that we compete at a national level, find some representatives. Like, in computer programming, I think we have four for each country, and then, and then, at the world level, each country sends four representatives to compete. So my high school senior made it pretty far. I think he made it to like top 15 in the country, and he felt bad that no one else from our school made it there, so he felt, oh, he wants to form a study group where he teaches more people in our school about coding so that more people can represent our school in that competition. And my friend wanted to join, so I tagged along because I thought, oh, coding, it would be fun. We would get to make some games and have fun. It's not about writing games at all, but I enjoyed it anyway. I ended up making it into the top 25 people [00:05:00] in that national competition program, and they gave me a quota for computer science engineering college admission in, in, in the top universities in Thailand. So I took that quota and, and yeah, and went into the computer engineering program. That's kind of how it started.
[00:05:18] Nicole Huesman: So you move from Thailand to the US, and you find yourself in Berkeley at University of California, Berkeley. Can you talk about how, how that unfolded? How you, how you ended up in Berkeley, how you ended up at Cal?
[00:05:35] Penporn Koanantakool: So in my third year undergrad, I did an internship at a national lab. The lab that picked me was the Parallel Computing Lab. They asked me, Hey, are you interested in writing code on PlayStation 3? So I thought again, Oh, is this about a game? That must be very fun. I'll get to put my hands on the new PlayStation 3. It was so new at the time. But it turns out, it's not about [00:06:00] gaming again. It's just like, Oh, let's just try to hack this new cool processor in this PlayStation 3 to run some scientific simulations. And it was really, really hard. It was my first time doing parallel programming, so it made me feel kind of like, Oh, I want to do it better. Yeah, I suck at this. I want to improve and I want to learn more about how to do it. So that's how I got into high performance computing. And then after graduating I was applying for a scholarship to go to study in the U.S., and I have to propose the field of study. So I was like, let's go with high performance computing. I already have some background in this and I'm interested in learning more about it. And that's how I applied to Berkeley. A professor I was close with, he graduated from Berkeley and he recommended I apply there, so I applied and that's how I got into Berkeley.
[00:06:54] Nicole Huesman: Nice. Very nice.
As we both know, many women who work in [00:07:00] male dominated fields like the tech industry experience challenges along the way, and I think this makes mentors and advocates really important to us. Can you talk about some of the mentors or inspirations that you've had along the way?
[00:07:21] Penporn Koanantakool: Yes, I have a lot, actually. So, we can go way back to when I was in middle school, I think. In fifth grade, my fifth grade math teacher, assistant professor Ravi Vanphomichit. She just told me, Oh, you seem to be good at math. Are you interested in competing in this math competition? So, I participated and I liked it and it motivated me to be good at math. I'm grateful that I was encouraged early in my, in my early years and, and I haven't been gaslighted or been said something that made me doubt in myself. Like I'm not good enough, or [00:08:00] maybe like I'm not supposed to be doing this or something like that.
It was strengthened by another teacher in my 10th grade. So, when I said there was a senior who did the computer programming study group, that teacher, assistant professor Dr. Sumali Ganjanashatri, she acted as the advisor of the project. And later on, we even thought of doing a creativity camp for middle schoolers that was all held by us high schoolers. And she was the main person who helped us make it into a real thing. She was a big inspiration for us.
When I said I applied to Berkeley, it's not just interning at the national labs. My advisor helped a lot too, and he is my other big mentor and inspiration. His name is assistant professor . He has a lot of insights in life and everything, and he is a really good strategist as well. So when I talk to him in my [00:09:00] undergrad years, I may understand like only 20 percent of what he said, but I thought, Oh yeah, that's clear. Everything. I understand it. And then when I look back, I was like, Oh, he did a lot of things for me. Like, this internship at the national lab gave me the background of high performance computing. So after I graduated, he also sent me to intern at a national lab in Japan to, this time, work on accelerating programs on the GPU graphic processing unit card, and that gave me more experience and I think it also helped a lot with my application, graduate school application, because the national lab in Japan is pretty revered in the high performance computing field.
And apart from Professor Puchong, there's another professor who helped me. I mentioned him earlier too, about applying to Berkeley. So his name is assistant professor Jitthath Phakjaanphon. Dr. Chittat is a really, really smart person. So he actually participated in [00:10:00] this computer Olympics competition himself when he was a high school student and he actually placed first in the world. So he was like a really big deal and a big inspiration for all of us, especially the students who have competed in this kind of program before. We all looked up to him. And I also met him when I was in the top 25 placing because we had some kind of intensive training camps and he was one of the professors who came to teach computer science. He was the person who suggested that I apply to Berkeley because he studied there before, and he thought I had the potential to be there, and he wrote me the recommendation letter. I think that also helped a lot too, because thinking of Berkeley's point of view, Gossipsat University is not a really well-known university from the U. S. standpoint. So having the experience in the Japan national lab that my advisor sent me to and also having a recommendation letter [00:11:00] by my advisor who also graduated from a U. S. university and also my Dr. Chittat who actually finished his Ph. D. Program from Berkeley and then he said he thought I could do well in the program. I think that helps a lot in my admission.
[00:11:15] Apart from my professors, there are also two of my close friends who I learned a lot about and get inspired by seeing them doing their everyday life things, but being so cool about it. Their names are Thachaporn Salanurak, or Ik, and the other is Maeng Por Pichayapothilimthana. Ik has a lot of influences on my choices as well. So I was having the mindset of maybe, Oh, it's a bad thing if someone asks you some questions and you can't answer. And then when I met him, and he, he's a really smart guy and, and, and top of class as well, and he just say, I don't know, whenever someone ask him a question that he doesn't know, and that make me feel like, that kind of like opened my eye. I adopted [00:12:00] his attitude, like why do I need to fret over the questions that I don't know, and pretend that I know about them, so I just say I don't know. I also joined the programming competition because he joined, I was following him. And we also ended up in the same university in undergrad. And I applied for the scholarship to the U. S. because he applied as well. He was a really great influence in my life and yeah, really great friend and role model in general.
[00:12:27] And Meng Po is also a really, really cool friend. She's very smart. She was actually the first female student to be the representative of Thailand to, to compete in, in, in the world competition. She ended up being my roommate at Berkeley, so we did a lot of things together and I, yeah, learned a lot of things from her and got inspired by her in everyday life.
[00:12:48] Nicole Huesman: Penporn, that's beautiful, the fact that you've had so many different, amazing mentors and inspirations along the way. I think that our [00:13:00] connections with other people are certainly something that, you know, it's so important and really shapes who we are and where we are in the world.
And I truly relate with the story that you told about the person who influenced you and changed your mindset around, you don't have to know everything because, for me, that was my father. He, and interestingly enough connections to back to Berkeley, but he was a physicist at Lawrence Berkeley lab. And one of the things that has always really you know, many things have stuck with me over the years, but one of them has been when you don't know something to say, I don't know, right? And to be confident about you don't have to know everything. And, you know, [00:14:00] interestingly enough, I'm actually really passing that along now to my son, right? That you don't have to know everything, right? Nobody knows everything. So that really resonates with me.
So you work in a really exciting area of tech that is just, you know, it's changing every day. And that is artificial intelligence and machine learning. And I wonder, what was your inspiration for working in that, you know, working in that area?
[00:14:35] Penporn Koanantakool: So when I was graduating, I was looking into what I wanted to work on, and it is similar to what I had been doing, which is accelerating scientific computations. I was thinking that it would be useful to apply my skill to help AI programs run more efficiently, be more environmentally friendly, consume less energy, something like [00:15:00] that. My other inspiration was from my aunt. My aunt had Parkinson's disease and she couldn't speak or write properly. It affects her ability to communicate with people. And AI wasn't even that mature at the time, so I would, I didn’t even know if it was possible to do it. I was thinking maybe I could use AI to kind of like recognize the patterns of how her eyes move to get some messages out of it. So I thought, okay, it's a good thing to go and start having some background in AI by working on AI software platforms.
[00:15:39] Nicole Huesman: There are so many possibilities and then risks or pitfalls with AI. You're working in it every day. Can you talk about what you see as some of the possibilities and then maybe some of the things we should be aware of?
[00:15:57] Penporn Koanantakool: Definitely. It is applicable [00:16:00] to pretty much everything in life and in fact, in our everyday life right now, we are also using a lot of it, whether we know it or not. If you search something, the website that gives you the results usually probably is using AI to rank the search results and find the most relevant results to you, or if you are doing some image recognition application or something like that, those also use AI.
And I'm excited to see how far we can apply AI's help into many, many more fields, like for example, in healthcare. Thinking of a doctor who has to look at some really, really big and elaborate MRI scan or x-ray scan of some people to find some fault in our disease in those images, AI can help highlight the areas that they should pay more attention to because it looks suspicious. And that would help a lot already. Or I can help diagnose more [00:17:00] things as well. For example, there's this one project where it can be used to diagnose or recognize diseases by scanning the retina image of people's eyes. And that helps, especially in the countries where it's hard to get a hold of the doctor or it may take too long to get to see the doctor. People get an earlier diagnosis of things that they may need to know about.
So yeah, I pay more attention to AI in healthcare, but it can be used to help other things as well, kind of like helping more in climate simulation, or other scientific simulations.
And now you also ask about fallbacks, right? Pitfalls. I think there are definitely at least a few. So the first one is fairness. There's a saying, you are what you eat, right? So, if you eat good and healthy food, then you will have a healthy body. Similarly, AI models are trained on a lot, a [00:18:00] massive amount of data gathered from the Internet. So the quality of the AI also depends on the quality of the data. So it gets tricky here because data on the internet, there is some biased data, there is some data that is not so good. So it really depends on the people who make this AI program on how they ensure to, and also represent everyone equally. For example, if we just feed more data of some specific race of people, for example, white people more, more than other races, it may be that that the AI program will do well on white people, but may not do as well on other races, or maybe not even recognize them. So It's hard to filter the massive amount of training data to make sure that we don't have too many biases or too many bad things that we don't want the AI to learn from, and also make sure to have equal representative of [00:19:00] everything and everyone just so that AI can be fair. This is a big field in AI called either AI fairness or responsible AI.
One other thing I can think of is privacy. Deep learning models are kind of like really, really big math models where we feed them data and then they learn how to predict things based on the feed data. So since it's a big math model, sometimes there could be some information that got leaked with the output. So there have been stories where people could find some queries to some AI model to make them spit out some of the data that it was training on as the output. And also because we need a lot of data to train the model, all our personal data is getting transferred more because thinking of people using mobile phone. Some applications might need the data of how you are using things or how you select your choices of watching a movie or whatever. Those get feedback to train this AI model so that, so that [00:20:00] they can make better choices for you in the future. So there's also a concern about, hey, what about if people intercept your data? So there are more chances that your data can get leaked.
I'll talk about one other thing, which is social equity. So AI makes it easier to do a lot of things now, so it can make people, so if you view it one way, okay, it can make people who may not have a lot of resources to actually get similar output to other people who may have more resources. But at the same time, not everyone can access AI, so it might also widen the gap as well. So, I think there are at least still billions of people who don't have access to the internet, or have a smartphone, and that would be their barrier, and the gap may widen because the people who have access can now do a lot more with the help of AI.
Nicole Huesman: With great opportunities come, you know, immense [00:21:00] responsibilities.
[00:21:00] Penporn Koanantakool: Whenever some new technologies come in, we have to adapt to it and ensure it gets used in the right way and so on.
Nicole Huesman: So Penporn, where you sit at this point, how would you define success for you?
Penporn Koanantakool: Right now, for me, success means being happy with where I am and also with what's to come, what I, that I see is coming. So as long as I'm content and happy, I think that's a success for me.
[00:21:29] Nicole Huesman: Looking back over your journey, what are some of the things that have been most rewarding for you?
[00:21:36] Penporn Koanantakool: I think there are two categories for the things that I accomplish as an individual and also the things that I accomplished as a team.
So for the individual thing, it may be being able to get the scholarship and come to study in the U.S. You need good grades and good extracurricular activities to be able to get [00:22:00] the Fulbright scholarship, and then also research experience with all the internships that I did to be able to get into UC Berkeley. So I think I'm, I'm proud of it because it's like a concerted effort of over a decade. I think that's one of my proudest moments.
And for the group thing, I think the one that made the most impact on me was that fun creative camp that we did as a high schooler, because it never occurred to me that high schoolers can just, like, just do this suddenly organizing a night camp. It's kind of like five days, four nights. And we, and all high schoolers did everything from start to finish. So, we booked the place where the middle schoolers would stay, which is kind of like one of the buildings in school. And we worked on all the documents and arranged the budget. So, it's [00:23:00] not free, we have to charge the parents who want to send their kids. So we did all the accounting and then the budget, prepared the food, designed the menu, and designed all the activities that the kids would do.
And, and, and I would say it's a very cool activity as well that I haven't seen at other summer camps before at that time and also even now I don't see similar things. We have fun activities that require team collaboration and also trade-off thinking and maybe applying some creativity in it as well. We have fun games like a quest, which is simulating an online game where you are a character in a big story where you have to go talk to other non-player characters just to get the info that you need and maybe some items that you need to make something to get to the goal that you want. That kind of online game was famous back then, so we made it in real life that the kids can just go and talk to the staff who would pretend to be like a non-player character and have items for [00:24:00] them and have some information that we will only say when the kids come with some specific questions or some items or those kind of things.
Or we have activities where they have to think of some trade-off skills. Like, a fun one and a classic one that we have and we used it in some of the camps afterward, it's kind of like catapult building. The kids would start with some certain fake amount of money and they have to use it to decide how much material they want to buy to build a tower and a catapult and materials such as toothpicks and rubber bands and those kinds of things. And so they have to decide on how much they would put in. And, and, and also kind of like the structure as well. How, yeah, what shape of your tower should it be? We would give a score to the highest tower as well. And for the catapult, after you are done building, you will use your catapult to attack other teams’ towers. [00:25:00] And the score would be like the highest tower at the end of the game. So if you build really high and really unstable, then someone else will bring it down and it will be just like a zero score if it just collapses on the floor. So, yeah, those kinds of fun trade-off things.
I was really, really proud of the camp and we actually won a national business planning award from the camp as well. And eventually after five years, my senior friend, the same person who actually organized the computer programming, was the founder of this camp as well. He spun it off to be a startup and it's still happening today. This year is the 20th year and the camp is still going.
[00:25:41] Nicole Huesman: That's awesome.
[00:25:43] Penporn Koanantakool: So yeah, that's being part of something cool.
[00:25:45] Nicole Huesman: Yeah. Yeah. That's fantastic. You know, and the kids get this fun experience, but they also learn, you know, it sounds like they also learn self sufficiency as well, right, and, and how to fend for themselves. And so [00:26:00] it's both, it's fun for them, but it's also so enriching and educational for, you know, and, and gives them life skills. So wow. I mean, that's incredible. That's incredible.
[00:26:15] Penporn Koanantakool: I also wanted to give a shout out to this high school senior, Mr. [name]. He is one of my inspirations as well and he also has an influence in my life, in my, how I made the choices in my life, because if he didn't hadn't started this computer programming study group, I wouldn't have competed in there and ended up here. And, yeah, and his organizing this camp start by starting this camp it let me learn a lot more about life skills and how to operate in general, with life in general, and it really widens my point of view, because I, before that point, I just mostly did academic things, just study and [00:27:00] compete in some, yeah, but I didn't really do activities, and that was the first time I got exposed to activities, and yeah, get to the fun of it.
[00:27:09] Nicole Huesman: Could you talk about some of the advice you might have for others?
[00:27:15] Penporn Koanantakool: I think my advice would be to surround yourself with good peers and mentors, kind of like your, they can be your safe cushion so that you can be more you can fall back to whenever anything happened to you, and they will also help guide you to the good way as well. Like a lot of the things that happened in my life were just some people telling me about it, or some people set an example for me, and when I have good people that I can fall back to, I, I feel like I don't have much to fear. I mean, what's the worst that can happen?
[00:27:47] Nicole Huesman: So that's a great segue into the question around what She Lights the Way means to you.
[00:27:56] Penporn Koanantakool: This relates to my personal experience a lot, too, because I get [00:28:00] inspirations or change my perspectives, or even sometimes change the way I think of what is possible and what is not by just seeing people around me, by just hearing stories. So I think She Lights The Way, it's a really great platform to let people hear more about other people's experiences and experiences and, yeah, get some food for thoughts and then grow together. So I'm really thankful for the opportunity to be here and share my stories and learn together with you, Nicole. Thank you so much. so
[00:28:32] Nicole Huesman: Thank you. The time with you today has been so energizing, and I want to thank you for really spending your time with us.