Open Book: Kushal Chakrabarti

Rob Williams

Published in

Open House

6 min readFeb 4, 2022

Open Book is a series of interviews where you get to meet our leadership team at Opendoor.

. . .

What do you do at Opendoor?

I’m the VP of Research & Data Science at Opendoor. My team and I are responsible for our core pricing and risk models, company-wide data science and analytics and data-driven breakthroughs. We have one purpose in doing all of this: make homeselling simpler, faster and more trustworthy for millions of Americans.

How did you first get into Data Science?

I can pinpoint the moment: ENG 77N, Introduction to Programming with MATLAB at UC Berkeley in 2001. That 3D contour plot on the MATLAB splash screen — it was like love at first sight.

Growing up, my teachers always used to say that I “had potential.” As everyone knows, that’s a bit of a back-handed compliment. I always knew I was smart-ish, but my brain was a bit different than others and I could never quite explain why it didn’t entirely fit what others expected of me.

Put another way, I saw (and still see) the world in numbers. And the moment I discovered the idea of programming and statistical modeling, it all just clicked into place.

I’m the sort of person that, if I get excited about something, I’m all in: I can’t not do it and I’ll blast through brick walls to make that thing happen. I finished my degrees in Electrical Engineering & Computer Science and Molecular & Cellular Biology in the first two years, and basically spent the latter half of what turned out to be my academic career doing full-time research. I took every graduate course in information theory, machine learning, statistics and data engineering I could find at UC Berkeley and published a few papers.

My plan at the time was to get a Ph.D. in machine learning and become a research professor. I really enjoyed research but Amazon ended up recruiting me in the early 2000s. From there, I turned down Ph.D. programs at Berkeley and MIT to start my first company. So, even though I never actually started a Ph.D. program, I like to think I have half a Ph.D.

You’ve worked inside Amazon in the heyday scaling years, been a founder and advisor to more than a few startups — and now Opendoor. How do you contextualize Opendoor vs. some of your previous companies?

Of the N=15–20 data science organizations I’ve seen the inside of, Opendoor actually most reminds me of early-2000s Amazon.

At the time, Amazon was a ~$25 billion dollar company and my team drove a few billion of that. Those aren’t accidental numbers. You don’t get to billions of dollars in revenue by accident; you have to be doing a lot of things right. But, when you peeked under the hood, there were a lot of heuristics. People use heuristics because they work, but they tend to work in the middle and fail catastrophically at the edges. My team was successful at Amazon because we were smart enough to see what the heuristics were trying to do, knowledgeable enough to map them to the relevant principled frameworks and effective enough to quickly implement and launch those principled solutions. Those solutions kept working in the middle but immediately started working at the edges, driving vertical-slope business wins for the company.

And that reminds me a lot of Opendoor today. In Q3 2021, Opendoor reported revenue of $2.3 billion. The quarter prior to that we generated $1.2 billion. Those aren’t accidental numbers. But, there are lots of heuristics behind the scenes. We’re unlocking the deep talent and experience across our teams to discover and implement those principled solutions throughout the company. We’re already starting to see those vertical-slope business wins and are looking for even more talent and experience to unlock more.

I understand you have a dog named Simba, a very fluffy and lovable German Shepherd. Tell me about how being a dog owner and training guide dogs has played a role in your life?

More than a few coworkers and friends have remarked that Simba is a bit like the concentrated form of me — playful, loyal and maybe a bit intense. I’ve been strictly a German Shepherd man my entire life, with one notable exception: I’ve trained three working guides over seven years for Guide Dogs of the Blind, all of whom are Labrador Retrievers. (Dusky, Lester and Louie live and work in Ohio, Washington and Pennsylvania.) Looking back, being a dog owner and trainer was a formative part of both my personal and professional life.

I got bullied a lot in middle school and high school and used to be a really socially anxious kid. As I grew up and took leadership positions at work, it took me awhile to realize that I wasn’t that same little kid who got punked around on the playground. By training, living and working every day with animals, I learned how to build confidence. Enough confidence to earn the trust of a 100-pound German Shepherd, and to eventually go on to lead and grow data science teams at billion-dollar companies.

However, my biggest takeaway was realizing that being the calm in the storm is an incredibly powerful leadership quality. Life can be chaotic and complicated — at work, at home, or in my case, raising guide dogs. Many young leaders get overwhelmed and start barking orders; they think they’re being assertive, but others just see the fear and indecision in their eyes. I learned the best way to handle tough situations is to be that calm, collected center of gravity for people during a chaotic time. By quietly and quickly assessing the situation and taking calm (yet decisive) steps, you not only maximize your chances of success in that immediate situation but also create emotional safety and trust. And that’s when something amazing happens.

Ultimately, research and data science is a creative activity: when you start a project, you rarely know what the right answer is — you have to play with it and connect points from lots of different places to figure out a good answer. One thing we know from decades of cognitive science is that you can’t be creative if you’re in a state of fear. When you get smart people, clarity of direction and an emotionally safe environment, that’s when the magic happens.

Simba being calm and watching over his 6-month old little sister.

What are other things someone might not know about you from your resume that you think are key parts of what make you, you?

My training and certification have long since lapsed, but I was a previously trained EMT. There was a time that I wasn’t sure if I would get into college, and I was looking for a respectable career. I figured I’d make a good paramedic and, to become a paramedic, you have to first be an EMT.

This isn’t just trivia. If you look at my career, it doesn’t follow the stereotypical path. Over the years, I’ve learned to reject the stereotype.

Last question — what are you most excited for this year?

I’m incredibly excited for the next stage of growth at Opendoor, both for the company and the Research & Data Science organization. The last couple years have been tough for a lot of us and it was for my team as well, but we did the hard work last year to get back to a healthy, strong place. The next few years will be a time for rapid innovation and we’re looking for the deep talent to help unlock even more of that.