Jason Dou founded Marbella LLC to teach AI thinking skills and reduce anxiety around AI-driven change. He studied at Peking University, Carnegie Mellon, and Harvard, and previously researched gerrymandering using Markov chains. The company aims to democratize AI skills rather than just teach technical tools.
Chapters
Okay, you're live.
Great. Great. Yeah. Hi, Jason. Yeah, hi, Matt. How are you? Good.
We're here today to talk about how you're gonna take over the world.
No, no, not taking the word. Oh. Our mission to use. Empower, democratize AI education. Empower our next generation who survive and thrive in the AI powers in power AI society. Yeah, that's.
That's what I meant.
My bad. Cool, Cool. Yeah.
So let's chat about your story.
Yeah. Because. Do I need to get closer to the microphone? Maybe. Maybe we need to get closer. Yeah.
Because your background's really interesting. You come from a family of family in academia. You're a third generation.
Oh, yes, student.
But also your dad and your grandpa were both educators.
Yeah.
Can you tell us about your story getting into AI and your. Your background and how it's evolved over the years and how you got to this point?
Yeah, yeah, yeah, totally, totally. Yeah. So as you said, my, my. My dad was a professor in a university in China. Yeah. So as I often say, really the professor parents like to send their kids to graduate school. That's how they really insist me to get a PhD. So I did my undergraduate degree in computer science from Peking University in China, which often referred to Harvard in China. It's a very competitive program. But I really have an American dream to study in the U.S. it's my first time came to the U.S. is I went to Carnegie Mellon University, which often refer as a top one in computer science, where Carnegie Mellon University and Peking University have a joint research program. So actually I worked over there. I work on a very interesting multidisciplinary research project called the Markov Chan approach to the Jerry Landing problem. So you know the gerrymandering problem, right?
Tell us about it.
Oh, sure. So gerrymandering problem is in that in the American voting system the party in power tend to manipulate the redistricting boundary to get unfair advantage in redistricting plan. So in that, in that time, I think the party in power in Pennsylvania was a Republican. They got like less than 50% of the votes but get more than 70% of the seats. So that's a German problem. So our idea initialized by a multi discipline team composed of both computer scientists, mathematician and historian and Samuel. So the goal is to use computer as a non partition tool to do the registrative distribution. So as a statistics here we think that after a certain like called the mixing time, then we can the algorithm control a reasonable radius plan from a uniform distribution. Uniform distribution means like every reasonable resisting plan have the Equal probability to be drawn. That's how we achieve this fair realistic thing. Yes. Okay.
Well we did just come for a month, so.
Yes, that's really I think little so the company where everything started, where Matt is an advisor called Marbella. Okay. Yeah. Huh.
That's great. That's great. But what about like. Let's touch on your story like back in what what did your like what led to you? What got you interested in AI and digital technology?
Oh that's, that's a perfect good question. So I, I, I think I was born to be pretty talented in mathematics. Yeah.
You got into Peking University.
Yeah, I did Olympic mass and a physics competition and won like top prizes and I had.
What was the top prize?
Like a first prize. What you get?
Did you get anything cool?
Yeah, first prize.
They give you something?
No, I think I got a certificate. Okay. I never gotten. So do you still have that certificate? Yeah hanging on your wall? Not hanging. I think the fall in my room. Yeah, my bedroom. Yes. I got a multiple one and the thing so why I was admitted like a discard admission without the college entrance exam. I think college entry exam is very famous in China but lots of people attended. But I got a got him before the exam so so that I can choose like among majors never like so I think I feel I have my talented in math and the physics is kind of limited. It's not like like that genius level. Then I noticed the school of Electronical Engineering and Computer Science which students there need to study some math, study some physics and also study something new which is information technology. So this actually did intrigue me. So I think I find I want to study that. So that's how I chose School of electronical computer science the ECS like just like course 6 at MIT. After our first year we started choosing a specific major. I choose the computer science because highly most relevant to mathematics. But I do have.
When you were studying, when you were at that point in your life do you like foresee the way that AI would evolve the way it has now?
Not, not really. Not really. So I think computer science and AI are highly relevant but still do some different difference. But at that time really the the area specifically interest me is the area called natural language Processing Computational linguistics. Actually now it's occupied not it's a huge part of artificial intelligence. You know AI is such a broad topic. Right. The logic like I say most application driven domain basically come with a vision talk about how computer process images and videos and another is natural language processing with how AI how computer process Language. So like the phenomenal application I ChatGPT. It starts actually like a chatbot. It's basically an application of nlp right. Natural language processing. So that's actually actually also where I started research in my sophomore year in college at Institute of Computational Linguistics at Peking University. Wow. So actually huh.
That, that reminds me. You have a lot of, a lot of research. You've done a lot of research in the industry. You have a lot of papers. How many do you know how many you you have? How many citations you have?
Oh my set, I have around 150 citations. 50. Whoa. Which. Which Some discipline is high, some eyes are finer. Yeah yeah. But I do I think that over years I did write a lot of paper part of. For PhD graduation. Yeah but I think so I really, I think to pursue research, really to pursue something you're truly passionate about. So I was doing the ARP research. It's really right. I discovered because in my first year in CS I studied a lot of algorithm which is pretty hard hard and you need to think of other corner case and, and which is kind of I think give me lots of pressure because you know I pick university. Many of my classmates, they start programming when they're nine or like even seven so they're already gold medalist. But I haven't learned any programming before I enter college. So I have lots of pressure in my first year also because I grew up have an interest in social science. Like I like news reading newspaper like international relations economy, like economics. So that's where I I found NLP or Natural language processing is where computer science interact with social science. So this like the perfectly match my interest and background.
It's amazing.
That's I think also why I love love the Harvard environment because as we all know Harvard is not the strongest in computer science but it has the perfect multi discipline environment. So that's why I feel I kind of survived the best.
Yeah. So you know this, this landscape of AI and the digitalization of the industry and the advancement of it. I mean it led you towards wanting to educate and wanting to empower and help people survive and thrive in this changing landscape.
Yeah. I mean what are.
Tell us more about Marbella LLC and what you're working on.
Perfect. Perfect. So okay so so I addressed this. So so mar. So the company inspired. So first I'll talk about the name of Marbella. So actually it's a name of apartment I used to live in New York City.
Yeah. It has a, a name of grandeur or like of relaxation and a nice, nice place to go, it sounds like.
Totally. So it's a, it's a, it's a high end tourist spot like a beach resort in Spain. So that's what I want to transfer out to. How to say transfer the message that for our generation of entrepreneurs, being an entrepreneur doesn't mean we need to live a very fragile lifestyle. If they still enjoy the relaxation and of Marbella. Be sound shy. We are being an entrepreneur.
Yeah, I know we were talking earlier and you mentioned as well that you know, it's approaching these, these changing times which come with a lot of anxiety and a lot of, you know, fear. Any sort of change that we have to deal with is really challenging sometimes. So part of that, that brand is also about approaching these topics which can be quite challenging and anxiety provoking and taking them in a relaxing kind of way of seeing the knowledge and information.
Exactly. I think. Exactly. I think recently when we chat meet with many professionals and student parents, they all have this anxiety and they worry a lot about AI. Right. They worry AI may take over their business, they worry AI may take their kids future job. But wherever you want to provide that AI is really a way to empower our work and empower our life to make us live a better life.
That's, that's great. I'm, I, I'm excited to start learning a little bit more, taking those courses myself so I can understand that a little bit more and live that lifestyle.
Yeah, yeah. I think one concept we try to try to educate HM is it's called AI thinking. So what do we mean by AI thinking? So this thinking, like the philosophy, the way of thinking is really come from our years of computer science, AI research and culture through a lot of research and project. So basically I think by AI thinking we want to teach our kids and students what's a part of a job can be taken over by AI. And we don't want to put out our labor work and repetitive work to do and where we can focus our true integrity, our genuine intelligence and innovation on the part AI cannot do. And that's the best take advantage of human capital so that we can do less work but be more productive and take advantage of the AI to do other boring work for us.
The AI can do the redundant tasks.
Exactly. You can leave.
That gives us a lot of space for innovation.
Exactly.
New things.
Exactly.
Moving forward.
Exactly.
That's great. So what do you do? So apart from business, you know, like.
Who, who do you look up to?
Who are your business idols?
Business idols or maybe not idols. That's the Wrong word. But who do you look up to in the business world? Oh, that's a good question. I think I admire many people. I think I mentioned that in my former podcast. So I think maybe as we all love Elon Musk who's an innovator and I was in a business school setting. I met multiple people who graduate from Harvard Business School. I think one is Steve Schwarzman who's the founder of Blackstone. Another is Ken Griffin who is a graduate of Harvard College and founded Citadel, the top hedge fund. I listened to their, their interview from David Rubinstein show. I remember Tim Griffin said when he was in college he spent a lot, a lot of time studying finance in the Baker Library, which is a B library. Yes. And then after I came to Har, I intentionally go to went to the Baker Library to try to trace Congress like a trajectory, you know. Wow. But then I didn't, I didn't find too many people study there. Time change. People don't study in library anymore. I know. Uhhuh.
It's funny.
Yeah.
How many amazing or how many, you know, known people spend a lot of time.
Yeah. In the libraries.
There's so much I don't know.
Oh, definitely. I think you remind me the tech entrepreneur. I seen a famous example at Bill Gates and Mark Zuckerberg.
Yeah, right. Everybody knows them.
Yeah.
You actually went to their dorm.
Exactly. I think this year, 2024, February 4th is a 20 year anniversary of Facebook. I did went to the Kirkland house, Zuckerberg Storm and also I noticed Bill Gates live in the Courier House also in that time. So living in the area really make us feel we are on the same page with these senior people and we want to create something as extraordinary as them.
Yeah. I think you know, we were walking on, you know, their legacies and everything they've built and probably most likely funded, you know, maybe even the studio so that, you know, we have all the tools necessary.
Yeah, exactly. Exactly.
Those worlds.
Yeah. Yeah. Cool.
I guess on a lighter note.
What.
Tell us, tell us, tell us who, who Jason is. What do you want most out of life?
Oh, oh, I want to definitely leave it. I. I said mentioned. I grew so I the play. I grew up all over China, but the place I spend the most time is a Hainan. It's like it's a southmost part of island in China. It's very similar to Miami. That's why when I come to Miami I feeling book at home. So I really love the beach, sunshine, lifestyle. That's also the reason I named my company Marbella. So I really want to have a.
Yes. Trying to bridge that world of digital technology and do it in a way that's relaxing and anxiety free.
Yeah, totally.
For education and moving forward into a world where we can focus on that innovation.
Yes.
And do it in a way that, you know.
Yes.
That works for our brains and not. Yes.
Totally. Well, that's really exciting.
Okay, well this is a good test run for today.
Wonderful.
Thanks for your time.
Yeah, thanks for coming all the way.
I know you flew in here on your I guess was it your dad flew from Japan. So just thank you for being here for Perfect Event and yeah, look forward to hearing more about Marbella llc. This has been the podcast at the Harvard Innovation Lab and yeah, we'll see you soon.
Yeah. Thank you so much. Thank you for having me. Yes. Perfect. Thank you. So I.
