Breaking the AI Mold with Jennifer Fu

Sheikh Shuvo: Hi, everyone. Welcome to the Humans of AI. I'm Sheikh. And here we're going to meet all the people behind the scenes who work hard to make the magic of AI possible. Today, we're talking with Jennifer Fu from Domino Data Labs. Jennifer, thank you so much for joining us. It's my pleasure.

Jennifer Fu: I watched your video. It's very good. Yeah, I really enjoyed it. I appreciate that.

Sheikh Shuvo: I appreciate that very much. The very first question I have for you, Jennifer, is if you had to describe your job to a five-year-old, what would you say?

Jennifer Fu: Well, for a five-year-old, I would say, oh, I'm a genie. I can make a dream come true. So you can have many dreams. For example, you want to think about ice cream. You can think about the color, taste, or even shape. So I can have an AI robot make it ready in five minutes. Or even like a car-shaped or something. Right? Yeah. Or maybe even less time. It can become true. So that's the magic of AI.

Sheikh Shuvo: After talking with a couple dozen folks, I think that's my favorite answer so far. Thank you. Well, Jennifer, tell us about your career story and how you landed where you are. Looking at your background, you've done so many different things across the tech world.

Jennifer Fu: Yes. Actually, many years ago, I started with AI, but it was ahead of its time. I was a researcher in computer vision at Fudan University, which is in China. And I wrote a lot of algorithms from scratch, right? Today you have a larger language model. Everything's just there. You take it, boom, boom, everything's done. But at that time, there was nothing. So you got very complex math formulas, converted them to algorithms. You write it and you see, oh, I have image processing there.

Sheikh Shuvo: That's a lot of hard work.

Jennifer Fu: Yes. Then you can see, you draw the contour of a face. It's not easy at all. Right. So you know, today you can use the iPhone. You unlock it with your face. It's so fast. It takes a second. You will never realize how much work it is.

Sheikh Shuvo: Absolutely.

Jennifer Fu: Yeah, it's a lot of work and things were not viable. So after a while, actually, I was a little disappointed. So I left. I wanted to do something real. So I joined an enterprise software company. I started building like storage, digital marketing, security in different areas. I worked at big companies like HPE and also after trying for a while, I decided to try different startups. I went to a number of them. Actually, two startups went very successful, the 3PAR and then Nimble Storage. Both of them went IPO and eventually, they were both acquired by HPE. So, well, I had a lot of fun doing this, and then I looked back, wow, the AI landscape is totally different. It has a lot of breakthroughs. Now you look at the foundation model, generated AI, what's incredible, incredible. Yeah, so now I spend more time. Well, I'm not going back academically, like a data scientist to build things, but I think I make it a platform. That's how I ended up in Domino Data Labs. We build a platform for data scientists to use AI products to produce things. I became a UI engineer, you know, leading the UI team to build those platforms. Now,

Sheikh Shuvo: For people who aren't familiar with UI engineering, can you tell us a bit more about what doing UI at an AI company means?

Jennifer Fu: Yeah. Doing UI is interesting, right? You have this foundation model. Like I say, hey, I want to draw a picture, like a style, the Picasso, right? How do you want to say it? The simple ones are just a field, right? You have a text box, you have a chatbot, you put a question, you get an answer. It's like chatbot stuff. That's called a simple user interface. You can make it more complicated. You can have more requirements like, oh yeah, I want to generate an image.

Jennifer Fu: I want a full image, a 10 image. I wanted the lighter and the darker or things you go to website, you build a thing, but the goal of the user interface is say, no user interface. You don't feel, you like talk to a robot, talk to a thing, you express what you want and you get your result.

Sheikh Shuvo: It disappears if you do it, right?

Jennifer Fu: Yes, you're not noticeable.

Sheikh Shuvo: Interesting. Interesting. Well, after so many different experiences of both working as a researcher, then on the user interfaces there, what are some of the misconceptions that you think people have about working in AI that you'd like to correct?

Jennifer Fu: Well, it's not to say you must AI company, right? You must, AI is everywhere. Hmm. That's a good point. Because also today, you know, AI is so convenient that like open AI has API, everything. You just call it one API, you get everything. It's so convenient. You don't have to be a data scientist to do AI. Mm hmm. Which I went to the 2023. A hardware summit, I was the only reactor engineer there. And people, you know, do a very complicated algorithm, right? So like a Tencent flow and a part touch, they do the, all those things. But yeah, I think it doesn't have to be like a high end difficult thing. Now it's a, it's everywhere, you know, so it can be any field everywhere. It's really, really, you know, make it, uh, anybody can use AI

Sheikh Shuvo: today. That's a great point there. Um, shifting gears a bit of especially coming back to your experiences as a master technologist at HP, then going back into, um, the startup world at Domino data lab. So what are some of the differences you've. Noticed in working on AI from an enterprise perspective and then in the, in the startup world,

Jennifer Fu: enterprise is a more formal. You have the process. It's good to learn this. So you have a guidance. Right. So, but also after you learn this, then you break all the rules, right? You learn the rules and you fall in order to break all the rules. Then the sky's the limit, right? You can try to invent, try and error, produce a lot of things. Well, it's not even imaginable today.

Sheikh Shuvo: Interesting. In your Day-to-day. Um, you're doing so many different, uh, aspects of, um, of ai, but what are some of the ways you use AI on a day-to-day basis in your own workflows?

Jennifer Fu: Well, we use the dominant data lab things, right? So we actually, we closely watch the marketer, what's the new thing? What's the, we'll build a platform. So this platform is the end to end, right? You can repeat your, your training things. You can look at a model, you can analyze it, all these things. But we want to see what's a new, you know, what's the data center wanted. Like for example, we have like a experiment management, right? So ML from experiment management. So we're building to our product. You have the model and it would build it into a dominant model. All this thing, so you keep watching what's new and what's your customer wanted and the beautiful platform, which is very easy for the custom to use that platform or the value for the AI

Sheikh Shuvo: and as someone who tries to stay on top of the latest and greatest in AI research coming out, what's an application of AI that you're excited about that you think?

Not enough people might be talking about it.

Jennifer Fu: Well, lately I studied some about the LensChain. It's built on top of the OpenAI. So it's a wrap on it, but you can build a chain. So kind of like a, you can chain, build many different models together. And then afterwards they have a LensMist. Then you can watch, you, they have the analysis of the result, the testing result, the training result.

Jennifer Fu: You can repeat it. You can see, oh, what's good, what's bad, but it's very encouraging.

Sheikh Shuvo: Do you do a lot of AI personal projects outside of work as well?

Jennifer Fu: Oh, yeah. I actually write for Medium. I'm writing about friendly technology. Oh, wonderful. Yeah. What's new in AI? And I try to mix them together. Not many people do that, right? But it's because of my experience. I actually learned... Yeah, many years ago, I was even teaching some AI classes. Then I used the language Perl. It's a very old language. Today, it's very different. So I'm very excited. Now, everything converges together.

Sheikh Shuvo: I'll make sure to include a link to your Medium blog in the comments here. It sounds like a great piece. My very last question for you, Jennifer, is looking back on your career, thinking about the type of advice you'd give to someone who's just starting out, say a software engineer just graduating college right now, who really wants to get involved in the world of AI but doesn't know where to start. What are some of the questions they should be asking themselves as they think about what company and what type of technology they want to work on?

Jennifer Fu: Well, you can start AI from anywhere, right? It doesn't have to be limited to enterprise or startup. But you know, it's good to start with something interesting, right? Try everything. Get your hands dirty in an unstructured environment. You might not get enough guidance, but you have no limitation to try and err. That's a great thing. And I think if you want to try, I would recommend a Coursera class, by Dr. Andrew Ng. He's my hero.

Sheikh Shuvo: Oh, of course.

Jennifer Fu: He's a standard of AI. He's a Stanford professor. He also founded a lot of companies, including Coursera. He has a machine learning course. I listened to a few of them. I think it's very good, very helpful. And I recommend people start with that. And then you can try different things. Don't be restricted to what's existing. Try what else we can do, right? Sometimes, just like a five-year-old, you can have the wildest dreams. That might be true.

Sheikh Shuvo: Yeah, it takes a lot of work to make that genie possible. Now, if one of the listeners wants to connect with you, Jennifer, what's the best way to find you online?

Jennifer Fu: Oh, you can go to LinkedIn.

Sheikh Shuvo: Okay, great. Awesome. Well, Jennifer, that's all I have. Thank you so much for sharing about your career and your life. This has been a great chat. Thank you.

Jennifer Fu: Thank you, Sheikh.