I grew up in a small town in Texas and am the oldest of four. My parents immigrated from India, so I was always balancing their culture with what everyone around me was doing. It was definitely challenging at times, but I’m grateful for my upbringing and having my siblings as lifelong best friends.
Many people I knew growing up went to a state school and then immediately got married and started a family. I knew I wanted something different, but with good schools like Texas A&M and UT Austin so close by and affordable, my parents would only help pay for a big move and pricey tuition if I got into a top-tier Ivy League school.
I ended up going to Stanford initially because I really loved the weather there! Stanford also has a really strong liberal arts program, and I wanted a well-rounded education. Instead of only studying engineering for four years, I also got to take several English classes and indulge my love of reading and writing.
The summer after my freshman year, I interned at Google. One day, we had a company all-hands with Doug Eck, a research director on the Google Brain team. He presented an open-source library for AI that generated music based on MIDI files. Given where the tech was at the time, it was a pretty basic demo, but I’d played piano and violin growing up and really enjoyed music, so it was a pivotal moment for me. I realized that with AI, I could unlock a slew of new tools to help all kinds of different people.
As soon as my internship ended, I went back to Stanford and signed up for as many AI and machine learning classes as I could. My first research project in the Stanford AI Lab involved generating music, specifically drum tracks, and determining if it was possible to intelligently fill in drum tracks for certain sections of different songs. The project wasn’t a runaway success, but it kick-started my interest in research. After that, I ended up doing several more research projects, collaborating with Google Brain, and eventually moving into the security and computer vision spaces.
When I was researching the implications and use cases of AI, deep learning was starting to gain real momentum. New language models and amazing new innovations were being developed, but I noticed there was a clear gap between the research and the practice of AI. So I joined a machine learning startup to build software tools that would help close that gap.
Before browsers existed, it was incredibly difficult to develop web apps because there was no centralized way for developers to build or interface with what they were creating. We’re in a similar place with machine learning right now, in that there’s no centralized way for people to deploy machine learning applications and interact with them. It’s basically a giant free-for-all until we can produce enough practical, centralized tools, and that’s really exciting to me.
I’m currently working as an entrepreneur-in-residence at Amplify on my open-source project called mltrace. I actually met the team a few years ago as an undergrad at Stanford, and I’ve always been impressed by how sharp and knowledgeable everyone is about developer tools, machine learning, and large-scale data analytics. Working with the team was a no-brainer for me, and I’m especially grateful for Mike and Sarah’s support.
I love to read and write. I’m also really into endurance training. In fact, I’m currently training for a half Ironman and hopefully a full Ironman soon if I can qualify. This fall, I’ll be starting my Ph.D. at Berkeley, where I’ll be researching the gap between machine learning production and research.