Jae Lee - Founder and CEO of Twelve Labs

(upbeat music) - Hey everyone, welcome to Asian Tech Leaders. The podcast where we interview some of the most interesting and inspiring Asian CEOs, entrepreneurs, and thinkers. I'm your host, Justin Pang, and I'm gonna mission to share the stories of Asian Tech Leaders to help guide your personal and professional life. Thanks so much for joining me, and I hope you enjoy the podcast. (upbeat music) - Jay Lee is a founder and CEO of 12 Labs, a video understanding platform that uses AI to help developers and companies search and understand videos quickly. Using its state-of-the-art multi-modal foundation model, it can accurately and instantly search exact moments within petabytes of video archives, generate coherent tech summaries, and perform prompt-based video generation. Prior to starting 12 Labs, Jay was an award-winning cybersecurity leader for the South Korean Army, and he previously spent time as a software engineer at both Samsung and Amazon. In this episode, you will learn where Jay is seeing the highest value from multi-modal Gen AI applications. Why Jay believed the multi-modal AI was the future and decided to start 12 Labs? And how building 12 Labs with a geographically diverse team is an edge. Hope you enjoy this episode, and let's get started. (upbeat music) (upbeat music) - AJ, welcome to the podcast. - Hi, Justin, thanks for having me. - I'm excited. - I know, we've been trying to get this on the books for probably three months, and I'm glad this time we're finally making it happen, even though we're in 2024, and you're actually my first podcast of the year, so thanks for kicking off the year. - Oh, I feel honored. - I'm sorry for the scheduling issues. - Oh, no. - I'm finishing up the, yeah, finishing up 2023 was a super busy (laughs) - Well, and you know, a lot of us in the tech industry, we know generative AI took over everybody's news feeds and mind share, and we can't walk away without hearing about generative AI, and they talked a lot about it at Davos back in January, which some people feel like that's when it's peak, right? Like when the top in Davos, and the world leaders ever talk about it, maybe that's a peak, but what's your take, right? You're in the middle of the generative AI industry, being in Bay Area, building 12 labs. Where do you think we are in kind of the timeframe of the potential for generative AI technology? - Yeah, I think that's a great question. I guess like before we get into that, so a little bit about 12 labs is, we are a company building video foundation models for enterprises and developers. So it's a multi-modal AI that kind of has human-like understanding of video content, and you know, developers can leverage our models for semantically searching for things within videos or creating rich summaries of video content or any video question answering type of tasks. And I think, you know, we are probably in like first innings or maybe 2022 and 2023 was like, period or like a prelude of the massive explosion we're about to experience. And in general, if in 2023, I think it was like a year of a lot of pilots and POCs, right? So usually with paradigm shifting technology, there's the initial spark of breakthrough, right? And then there's enterprises that are really shocked at the capabilities that this new technology brings and that they go into this phase of, okay, we're gonna try out every single use cases that we just couldn't solve, right? So hence the year of POCs and pilots. So I think some of those are less mission critical, some of them are mission critical. But I think, you know, even with all the buzz and hype around Gen AI, you know, we haven't gotten into addressing super critical use cases yet, you know, given how, you know, language models are being used for right now, like enterprise knowledge search, right, with REG. And still the copyrighting and code gen is, you know, still the most popular use cases amongst many other, but I think in 2024 and 2025, we'll see really awesome use cases, like mission critical use cases being addressed with this technology. So we're very early in the innings, right? And I think we'll still see the trend of, okay, we don't know how far these models can go. So let's, let's green out more capabilities, right? By training maybe larger models or using more compute, but then we'll start addressing the issues around cost, you know, making them more efficient, smaller. Yeah, I think we're gonna see that trend as well. - Yeah, so still, well, I guess number one, a lot of capital has obviously been deployed because to your point, you know, the computation costs are sky high and you need that kind of as a baseline. Going back to kind of your point on the value, at least for 12 labs, focus on, you know, better understanding video and extracting value there. What industries or types of enterprises do you see having the most initial success with a solution like 12 labs? - Yeah, so I think in 2023, we've been reached out by hundreds of customers across 10 plus different industries, right? So I think, you know, for an early technology like ours, which is called video language model, I think it's incredibly important because we don't know how capable this model can get. It's very important that we are strategically aligned with our customers in terms of our research and product roadmap, right? So we saw a lot of extraction in media entertainment. They have a lot of content that they, you know, video understanding is like a hair on fire problem before then, right? Because, you know, not only, you know, there's the end product or end films that we watch, but there are a bunch of amazing D-rolls or raw footage that could potentially get used for, you know, further monetization, right? But in order for you to do that, something you need to be able to find granular moments, what's, you know, and current technologies have hard time kind of empowering the enterprise in media entertainment to really address that issue. The other one is contextual advertising, right? So without leveraging like your personal information or cash information, how do you serve contextually relevant ads by looking at the content that the viewer is watching, right? So finding right, you know, and insertion points through contextual understanding of videos is a big opportunity and we have a couple large companies and smaller companies that are trying to address this. And yeah, and we have a lot of fun helping them. And then the last is, you know, public safety, right? So helping, you know, the folks that are monitoring, you know, surveillance footage for public safety, you know, helping them find critical moments as well as give them the power to get real-time alerts on certain kind of behaviors that they would just register via text, right? And we'll be able to kind of do the surveilling for you. So you don't have to watch like 500 concurrent camera feed. - Right, right, yeah. - Right, yeah. And you know, even to the first use case you mentioned, that's something I wish existed more mainstream, like even within, you know, large video platforms like YouTube or if I'm watching a show on Netflix, for example, I'm looking for a specific part of, let's say, an interview on CNBC. I'm kind of blindly just moving forward the time lapse scroll to figure out where that point is. And it's not very precise. And it could be way more efficient, right? If there's a way to just do that semantically and-- - Yeah. - Obviously so. - Right, if there's a system that can like understand the context behind the content of all your videos, then you now have this capability to build up really engaging like fan platforms, right? You'll be able to have agents that your viewers can interact with and it'll be able to pull the right content, to be right highlights. So yeah, in that sense, you know, I think we are having a lot of fun building this bedrock in infrastructure to make that happen. - Yeah, so can you kind of rewind a little bit? I think 2021 was when you co-founded 12 laps, is that right? - Yeah, yeah. So I started 12 laps, Mark of 2021 with four other co-founders. Four dudes won't do that. I, yes, to do that, so Young is our head of go-to-market. So and Shin and I go back way back actually, back in '09 that when we went to prep school together in New England. So, you know, we've kept in touch and she loves, she loves talking to developers and technology in general. Right, and then the dudes met at the Korean Cyber Command. So I was born in Seoul, grew up in Knoxville, Tennessee, and then went to prep school in New Hampshire, and then, you know, Berkeley for computer science, but I, you know, wanted to retain my Korean citizenship. And I was fortunate enough to have, serve the country with keyboard, rather than a right to go. And the rest of the technical co-founders had similar background, right? So we've done a lot of video understanding, like multi-modal video understanding research for the Korean Cyber Command. It's very similar, it's modeled after, like the Israeli estimated 80 to 100 or US Cyber Command. Yeah, so we have always been kind of very much into building like intelligent systems, right? And the co-founders, you know, we were thinking of pursuing the Korean academia and going into research, but then the kind of research and product building that we wanted to do required scale, right? And we had a lot of fun working together. So that's when we decided to hate. That's built something of our own and see if we can make some breakthroughs and along the way, maybe we'll be able to find great partners that can support us and continue building, that's how it all started. - Was that a difficult decision for you personally of like deciding to venture out and build something as opposed to work at a big tech company, right? You've done internships at Samsung and Amazon. So like, what was that mental calculus for you at the time of which direction you wanted to go? - Yeah, so I think, I mean, starting a startup is always hard, but then, you know, I had this idea that like maybe not doing this is like, maybe like, it's a bit too extreme, but I had this feeling that, you know, a lot of startups fail even before they do anything because of like this co-founders issues. I have like, amazing, like my best friends basically, and we have a lot of fun working together. So it seems like we've kind of reduced the uncertainties around-- - You've reversed it if on their dispute. - Yeah, and then I was thinking, hey, I'm like still very, very young, and it's probably like, I'm not doing any favor to future Jay if I don't, you know, pursue it, right? So when I had that kind of, you know, mindset, it was like pretty, pretty simple. It was like, hey, we're gonna do it. - And was the idea more around, let's build something interesting that works, or at the very beginning, did you have this idea of like, I really want this to be a company and raise money, and like, how big was the vision at the time of the founding? - Yeah, so we always knew from the start that multimodal AI is going to be the next kind of frontier, right, we've already seen. I think a lot of people have felt that when the attention is only paper came out, and you know, starting with like 2018, like we saw companies like OpenAI scaling language models, and we've seen some traction in, you know, image models as well, right? And what we thought was like, we need to be able to, kind of build a system that can train on the kind of data that resembles the kind of sensory input that humans kind of form their, you know, representation of the world around them, right? And, you know, video felt like the kind of data that resembles the sensory input that little kids kind of, you know, see or feel, right? Even before they develop the language skills and text-based reasoning skills, right? And then, you know, this whole field of like multimodal AI and video understanding was still very, very nascent. So we at this feeling that maybe your capital isn't a mode yet because there's so much underlying research to be done, right? So our bet was a company that raised $100 million versus a company that raised maybe like $10 million. But this smaller company has stellar team that can kind of reduce the uncertainties around the fundamental research might have better chance, right? So that's how, you know, we got started. And the vision of the company has always been, you know, we're going to build the semantic encoder of the world for all future agents that need to understand the world like we do, right? So that vision kind of trickles down to our research ethos and our video first approach. - Very cool. And then the other thing, you know, I found fascinating and also inspiring about 12 labs is a good proportion of your team is actually based in Korea, right? And obviously you're based in the Bay Area. Did you have any other companies or, you know, role models that you looked up to that actually you had this model of, you know, part of the founding team is in Asia and the other part is in the Bay Area or did you feel like you were one of the firsts? - Yeah, I mean, there's definitely a few companies that have similar structure. So Sandbird being one of them, it's there are the largest chat API company. John, the CEO at Sandbird is also one of our, you know, board members, right? So ever since I was little, the kind of company that he was building was inspiring. But also I think, you know, the founders are need to be kind of positioned to kind of take advantage of our kind of bilingualness, like understanding of two cultures, to marry the best of Korea and best of Bay Area, right? And build a diverse, really creative team. I think there's a lot of unique and diverse group of people here, but then like the Bay Area could also kind of feel too homogeneous sometimes, right? So for me, I always strive to like expand my perspectives and I think having offices in both Seoul and the Bay Area kind of brings in a lot of different perspectives for us to build a technology that is very new, right? And it requires having a broad perspective as well as that person, I think. - So definitely has been an advantage, right? Like helps you think out of the kind of homogeneous monoculture that might be the Bay Area and-- - Sure, for sure, and-- - Yeah, yeah, absolutely. And the talent that we have in both offices is incredible too. So yeah, it's been great. And I'm kind of the topic of talent and support. Your investor base is like the who's who of AI luminaries, like Bay Bay Lee, you know, the co-founders of co-here, we had Ivan on the show previously. Alexander Wang, you know, a bunch of, a bunch of very influential and leaders in the AI space. What are some things that you've kind of garnered from the advice that you've gotten over the years around building your company and where you should be focusing your energy and attention? Any common patterns that you've seen? - Yeah, so I think we're really like fortunate to have the backing of these people, right? And I think the, the competition and the kind of company we're building like deeply resonates with, you know, Bay Bay is, you know, she's spent her entire career in building machines that can, you know, see the world and understand the world like humans do. And, and from the advice that we've gotten, I think probably the most important one is always making sure that we are solving a problem that is, that has incredibly high impact in both research and, like, economically, right? So, you know, not chasing after hype, right? We've seen explosion in language models. And I feel like some founders might have been like, "Oh, do we train a language model of our own?" But I think, you know, being incredibly focused on solving this problem of video understanding, which is a very important, and one of the hardest, I think, problems in modern AI. Because if you had solved video understanding, that would be you figured out the right model architecture and learning algorithms to leverage the world's most redundant data, which addresses the issue of data shortage right now, when we're seeing, right? And also, given, you know, video accounts for 80, 90% of the world's data, like, it's kind of without a question you need to understand all of this data, right? And have a system that can learn from it. So, there's definitely a huge economical benefit to it. So, knowing that it's early, but, you know, putting all our efforts into solving this problem and pushing the boundaries as fast as we can, it's has always been the most helpful advice, right? Yeah, right. And even the, you know, the foundation model that you're building, to the extent you can share, what are some examples of the corpus of video data that you're actually using to train the models? How much of it roughly is public front in the public domain versus private? Could you share any information on that? Yeah, so I think the way we approach data is quite sophisticated. I'm really happy to have partnered up with radical ventures where, you know, these guys are incredibly serious about building the next, you know, you know, generational AI companies. And from the start, we had AI ethics and data policy that our companies to follow, right? So, under the guidance of our legal counsel and board members, we gather publicly available data as well as proprietary data that our early design partners and customers are willing to provide, because they see so much value in having this model built for their business, right? And of course, we'll, you know, those fine-tuned models are, you know, theirs to leverage and use, but yeah. Great. Now, I would love to like rewind a little bit even further, which is your upbringing. So you kind of mentioned, you know, you had have brain citizenship, but you spent some of your later years in the US. Can you start at the beginning, where were you born, what were your parents like? Yeah, I was born in Seoul, but I spent most of my time here in the States. My parents are super hardworking business people. They both owned their own businesses. My dad was running like factory optimization, kind of technology company for steel industry. And my mother was running like a promotional goods and services company, right? And, you know, they were incredibly busy, but I think the lessons that they gave me has always been like responsibility and then freedom, right? So, you know, my parents weren't like, typical kind of like Asian tiger parents where hey, you have to like study and get all these scores, but I think they put more focus around decency and honesty and transparency and making sure that, you know, I have fun growing up. (laughs) So, you know, just around being an honest person, like growing up as an honest person has always been like their focus. And then, you know, having the four to two, kind of do what you want to pursue, right? And that's why like I decided to rot in my perspective, that was like around fourth grade. I was like, "Huh, I want to kind of explore a new world." And at the time by uncle who is still, you know, serving the Korean military as an officer was offered to get his PhD in University of Tennessee in statistics, right? So, I've tagged along and ever since then. I guess my uncle was a bit like a tiger military, kind of like a family member, right? So, that's when I got exposed to a lot of statistical learning methods, some of the OG methods that statisticians had put together to make sense of the world, right? But yeah. - What age was that when you tagged along with your uncle? - I think I was 10 or 11, oh wow. - Yeah, yeah. - So, did you actually, like geographically move, or would you just like? - Yeah, yeah. - Oh wow. - Physically moved. So I finished my elementary school in Knoxville and middle school, and it took my uncle, I think four and a half years, five years to finish his PhD. And then, you know, he was going back to Korea to become a professor at the military academy. And then, you know, I was like, do I go back or do I stay? And how do I stay, right? When there's no family members in the US. And, you know, that's when I kind of so learned about boarding schools. - Oh wow, wow. And were your parents pretty supportive of that? - One, two. - Yeah, I mean, my parents were worried, right? It's like, yeah. If this is what you want to do, work hard, and, you know, if you can get in, you have to get in the first place. - So, again, let me get that, like around 10 to 14 or 15, you're living with uncle, right? - Yeah. - But you moved to the US, is semi here, obviously being taken care of. But your parents were in around, and then around 15, 16, is that when you did boarding school? - Yeah. - So you've been independent for a while? - Yeah, I think my parents kind of missed out on all of my, you know, teenage years, very pretty, which is like good for them, I think. - Absolutely. - They still have a very pristine image of like a nine, 10-year-old Jay in their head. - Yeah, but they would visit me frequently. But, yeah. - Wow. And then, what were you curious about when you were younger, like, you know, in the 10-year-old or before 10, what you talked about optimizing for fun, like having fun, you know, being in on this person, were there activities that you remember fondly from your early childhood? - Yeah, I think, you know, back in like, I guess elementary school, didn't study at all. I think I was really interested in just learning about people and just hanging out with like a bunch of different groups, just fascinating to see how differently you can think from somebody else, right? Just learning about their perspectives and why would that person do this thing in this way, right? So I guess as long as I remember, I have always been very curious about understanding the surroundings and the people in it. And with that, I think I had this like, more of a scientific influence from my uncle where, you know, he was doing a lot of Bayesian statistics back then and I guess when I was 11, was when I kind of got the first introduction of distributions and, you know, even though people may seem very, very different and their perspectives are very different, you know, they're all kind of part of distribution, right? And then you will sometimes see an outlier and you probably want to be that outlier, right? And the idea that you can capture, you can have an algorithm or system that can learn through data and have like this distribution learned, like this complex function of our life and world was always fascinating. So yeah, so I guess deep down it, you know, comes from, and 12 leaps to comes from like my kind of interest and wanting to understand different perspectives and people. - And then in university, you studied computer science, right? Was that a easy decision for you in terms of what program to study or how did you work? - Yeah, yeah, I mean, I had always been, so I started coding like, I think sixth grade, like I was just copying my uncle, like writing code in MetLab just, you know, because when you see an adult man that you look up to doing something for like five, six, seven, eight hours a day, kind of like sparks some interest, I think. So I had always been very much interested in wanting to learn and be better at coding. So I keep down on a lot of sciences and computer science back in Exeter are back in high school and study some more computer science in Berkeley. And the whole New England kind of preppiness, I don't think I've really enjoyed. So I want to, like, this different end spec from and extremely of, yeah, extreme hip in this and just like being free. - Yeah, yeah, they had a good experience at Berkeley and then during your time at Berkeley, that's when you did your internships, right? At San Sang and Amazon. - Yeah, so I think that's when like, you know, the pure pressure is kind of come in, right? You know, I had always wanted to research, but then I think, you know, back in like 2013 through 20, like 16, 17 was when like there was this crazy growth, I think. And, you know, I was talking to a few of the classmates that are few years older and they're like, don't do research, go to industry and build something cool, right? So that's when, you know, I decided to kind of, you know, have an internship at Amazon, which was growing at an incredibly fast pace back then. And then San Sang, well, I was born in Seoul, so I, but I didn't get to spend too much time in the country and wanted to kind of see what it's like to work there and kind of have experienced the world culture of one of the largest, you know, companies that came out of Korea. So, yeah, so those two kind of roles were for me to, like gain maybe like what it's like to work as a software engineer in the industries. So it was good learning. I had a lot of fun building cool things. - Are there any lessons that you took for, again, I know it's an internship, so flew by. - Yeah. - And the lessons that you took from that experience that you still think about and apply today, or is it seems so far away in like a different world that it's hard to connect the dots? - Yeah, I think the lessons that I learned was never settled. (laughing) It kind of felt like, yeah, it kind of felt like, because these companies are so big and it's, there's like incredible processes that put it to place for you to focus on this one specific thing out of like this big puzzle and it still kind of keeps the train running, right? So, you know, that is incredible, but I think in terms of learning your perspective and wanting to build bigger, faster, probably less ideal, right? And I think I had to tell that deeply when I was working with those two giant companies, right? - Yeah. - Yeah. - And then knowing that I could probably come back after I've honed in my skills and I've done the kind of things that I wanted to do because these companies will probably never go away. (laughing) - That's your safety net, right? You're insurance in case you ever want to go back. - Being a founder and CEO, right? It's a lot of context switching, many different stakeholders who are equally important. How do you manage your day to day? Like specifically what some might feel is a lot of pressure and at high expectations. Whether it's investors, employees, other stakeholders, how do you manage that? And is there anything that you've been doing as part of your daily or weekly practice that you found helpful? - Yeah, I think there's no silver bullet, but probably the most important thing is, I think the most important job of the CEO is relationship management and relationship building because at the end of the day, you know, 12 labs, like the company is built by people, right? And I think the best way for you to manage relationships and nurture them is by listening a lot. So I try to spend a lot of time listening to everybody, right? And so we have this thing where, you know, I meet with a lot of my members with one-on-ones and I try to listen a lot and then formulate and there's a lot of like jewels in things that people say that kind of, maybe it's not applicable right now, but having heard different perspectives and it's probably like stored in your subconscious, right? In your subconsciousness, but then once there's like this big decision to make it kind of all kind of, you know, matches up, right? So I try to listen a lot. So on a weekly basis, I have like a day where I'm spending a lot of time doing one-on-ones. But then, you know, stress management is also very, very important, but maybe this is not for all founders, but I think ever since I started 12 labs, I had just, I listened to a couple mentors and I just kind of like make a promise to myself, hey, 12 labs is probably going to be your life for the next 10 years. So if you're okay with that, start. If you're not, if you think like in the neck, in the, I don't know, three, four years into the journey, if you like, if you want to like, I don't know, go to long vacations or do something else then don't even start, right? So having that mindset also helped out, right? And also, you know, it's been very, very fun, actually, right? Not knowing that like this is kind of like your life and almost like your baby too and you're nurturing. So yeah, kind of makes you understand your parents better too almost, it feels weird, right? - Yeah, I guess we're looping back to your parents. Do they have an idea what you do? Like, you know, even, you know, people's spouses don't necessarily really know what their spouse says, but, you know, when you talk to your parents or when your parents tell other family members about what Jay's up to, do you feel like that we could sense of what you're building and doing or not really? - I think they understand it now and they appreciate it, but when I first started, my dad was like, are you building an app? Like a, like a calculator app, but yeah, well, now that 12 laps has made some progress and then, you know, we've popped up in New South Lids and things like that, you know, I think they spend some time learning about what their son is up to. So yeah. - It's fun. Yeah, no, I played around with the 12 laps platform the other day and it was great. Like, I'm like, I need this for all of my videos, both like in my phone and my personal photos and videos with my kids and also the shows that I watch. So it's an amazing product. - Thank you so much. Yeah, it's the, it's the first kind of model that we've released, but we're hoping to release new model like on a monthly basis now. So you'll probably see it improve rapidly. Yeah. - Yeah, great. And then, you know, to close off, what are you most excited about for the area? We're one month into 2024, but if we were to reconnect in December, what are you hoping to have seen or achieved or have done? - Yeah, those are, yeah. I mean, there are so many that I want to achieve, but I think the most important thing here is building great team and maintaining our company culture. Right. So I think 12 laps has moved incredibly fast and has gotten to where we're still very early, but has gotten to where we are. I think mainly because we have a company culture that appreciates diversity, creativity and decency, right? So, but when the company is growing really fast, it's easier for you to kind of lower the bar, right? Because you have to grow fast, right? So maintaining that culture and having a fluid culture of marrying science, engineering and business, even when we are at 80 plus people company is very, very important, right? And making sure everyone's still aligned on the vision and mission of the company. Once you have that, I think like good things just happen, right? - Yeah, 'cause you have like the right people on the quote unquote bus, and then you feel like you can tackle any problem, right? When there's a high level of trust, you believe in each other, you compliment each other as well, right? You're not hiring for one skill set or one culture type. I guess on that point, do you have any favorite interview questions that you're comfortable sharing, right? Like not to bring up to anybody who ever interviews with you, but are there one or two killer questions that you typically like to go to, to help answer the culture question? 'Cause it's really hard in a series of one hour meetings to figure out, is that person gonna be a fit? - Yeah. I mean, we have a dedicated like culture interview, like CEO culture interview, so I ask a lot of questions, but I think my favorite question to ask is, how would your colleagues from your previous company, how would they describe you as a colleague and as a person? And would you agree? And how do you see yourself, right? And is there a big discrepancy? - Right. - If there is, why? Well, you know, and then I also ask like, you know, what their superpowers are and what are some of the things that you've worked really, really hard but failed, right? And how did you cope with the failure, right? Yeah, things like that. - Yeah, great question to understand resilience and being able to bounce back, right? And honestly, some of the times the best answers are not work related, like there are failures that can happen in our personal life where you're like, well, I really wanted this husband, it didn't, and it's sought, and this is what I learned from it. - Yeah, but I think, you know, there isn't like a single killer question, but the series of questions will, I think ultimately what I wanna know is, do I wanna spend, you know, having talked to this person or like an hour or an hour and a half? - Or would you wanna go on a cross-specific potion flight with them? - Yeah, yeah, do I wanna spend like 16, 17 hours with him like sitting next to him? - Exactly, or heart, yeah, that's very important. - Thank you, Jay. This is a lot of fun. I appreciate the time in there, you're super busy. It is great to connect, great to hear your story. Really psyched to see what 12 labs continues to shift, and we're all rooting for you. Where is the best place for folks to find you if they wanna follow you on the internet? - Yeah, yeah, I'm not on the internet too often, but LinkedIn and I'm trying to up my Twitter games. - LinkedIn, okay, good. You gotta, you gotta become an AI influencer as well, right? So, yeah, yeah. Hopefully, hopefully our work at 12 labs speaks for ourselves. - It'll go viral. - It'll go viral at least, yeah, yeah. - Awesome, thanks so much, Jay, I appreciate it. Take care. - Thanks so much, Justin. - Thank you. - I had a lot of fun. - Thanks so much. - Bye. (upbeat music) - Hey everyone, thanks so much for listening to this episode of Asian Tech Leaders. If you enjoyed the podcast, please share it with your family and friends. Leave me a review on iTunes, or drop me a note on our website, AsianTechLeaders.com. I really appreciate having each of you as a listener and sharing your valuable time with me. Be well, stay healthy, and follow your heart. See you soon.