September 07, 2018
Plaid People: Kevin Hu
Updated on November 19, 2019
A data scientist at Plaid since March 2017, Kevin Hu is busy pursuing his dream of building things that matter. Here, he talks math contests, cats of Instagram, and the collaborative nirvana that is “group sound.”
What exactly does a data scientist do all day?
I help build models that try to find meaning in data. For example, if you think about words, there is some intuitive concept of distance, of how close or far words are in meaning. Synonyms are very close; antonyms are very far. Then there are words that are completely unrelated, which puts them in a totally different dimension. That can actually be modeled in a given context—say, financial transaction descriptions—and from the models, we can predict meanings of aggregate phrases and sentences. That ends up helping consumers set budgets and identify what their expenses are going toward.
At Plaid, every project that I work on could be someone’s dream data science hackathon project. There are just so many different angles to explore and so many ways that we can build things that could be really impactful.
When you were a kid, did you dream of becoming a fintech data scientist, or did it just work out that way?
I studied applied mathematics at Harvard, and before that I competed in math contests in middle school and high school. The finance world and the math competition world are very heavily intertwined; a lot of finance companies sponsor competitions, so I had heard about all of them. It seemed like a very standard path for people who did well in math competitions to end up going into finance.
And that led you to Plaid?
Well, first I interned at a trading company in Chicago and then at Goldman Sachs in New York. Those experiences really helped me figure out what I enjoy working on, like modeling projects, building actual products, and working with people to figure out what I could do to help them. After Harvard, I was a quantitative developer at a high-frequency trading startup in Chicago, and then I took a position at a startup in San Francisco. That was my first foray into the tech world.
And that led you to Plaid?
Yep. After I moved to San Francisco, I went on a hike with my very close friend, Joy, whom I had met many years earlier through math camp, and she vaguely told me about some of the challenges the company she worked for, Plaid, was working on at the time. And then she mentioned that there were data opportunities.
What was it about Plaid that appealed to you?
As I thought a bit about the volume and richness of the data that Plaid has, it became super appealing to me. It got me thinking that Plaid has the best chance at democratizing finance out of any organization around. Given the company’s incredible dataset and also a lot of the recent advances in machine learning, Plaid has a really great opportunity—almost a responsibility—to help build tools that help everyone with their finances, not just the people who already have their wealth managed by banks, and to push finance as an industry toward making everyone better off. The potential for that impact really appealed to me.
Now that you’re a more than a year in, what’s your favorite workspace at Plaid headquarters?
I tend to be somewhat nomadic in terms of where I work, but I really love being in situation rooms where people are working on a really challenging, urgent problem. I think that’s because you really get to witness the ingenuity of the people you’re working with. And I think the common effort and thought people put in brings teams closer together.
So, you enjoy working in teams. Is a lot of your role collaborative?
I work with teams as much as I can because I think that figuring out what our clients are looking for in terms of what our data science team is building, is really important. Client feedback, both internal and external client feedback, is really critical to figuring out what to prioritize and focus on. We can always go more advanced and get more accuracy and try to build the best data science products, but we want to make sure we’re being practical, and solutions oriented, and really making sure that we’re addressing as many of the client needs as possible.
Rumor has it you double as Plaid’s unofficial/official staff photographer. How did that come about?
I actually became interested in photography after I started at Plaid. A lot of people here are awesome photographers, and they enjoy talking about the technical and artistic aspects of it. I learned from them.
I really like capturing the way people look at things, and that lends itself well to Plaid events. I also like capturing animals’ emotions—the way a coworker’s dog looks at his human with complete love in his eyes, or the way my cat looks at food that I’m eating, because he always seems very emotional about it.
So, you’re an animal guy?
Completely. I’ve always really loved dogs and cats, and I’ve also had some interest in medicine and behavioral sciences. So, now, I volunteer in the shelter medicine program at the San Francisco SPCA every Saturday afternoon. I mostly administer medication to companion animals and work with physical therapy cases.
Do you have a hard time not adopting every animal you work with?
So, I adopted my cat Finn from the SFSPCA earlier this year. He’s awesome. He’s super sociable, he’s very clingy, and he’s on Instagram at @catfinnated.
Amazing. How else do you spend your time outside of work?
I have been playing the violin since I was a kid. And really, any chance I can, I like to play, whether it’s by myself or, better yet, in small groups. In college I played a lot of trios and quintets, and also performed in an octet internationally and for President Obama. There’s a concept that I really like to work toward called group sound. That’s like a Holy Grail that groups will spend a ton of time trying to feel out. Obviously, when these groups are playing music together, there’s some level of individual contribution. You want to get your notes right, you want to play the correct rhythm, everyone needs to actually play together. But then, at a certain point, the group figures out a central sound and idea, and everyone is working toward that together. And when it happens, it is really striking and really beautiful to witness.