सम्यग्दर्शनज्ञानचारित्राणि मोक्षमार्गः।
Tattvārtha Sūtra 1.1
[Right [world-view, knowledge, conduct]] is the path to liberation.
I’m a second-year Computer Science PhD student at Columbia University. My research focuses on continual learning and reasoning.
Outside my research, I get interested in pretty much anything. I love to learn and take notes as I go. I’m also trying to write more.
Before Columbia, I trained LLMs at AWS. And before that I studied at Carnegie Mellon and BITS Hyderabad. I’m very grateful for all the amazing people I’ve met along the way.
I’m currently in New York and grew up in Mumbai.
Please reach out if you’d like to chat about anything or if I can help you in any way. The best way is through email.
Now
Last updated: October 5, 2026
I’m thinking about:
- Measuring interestingness of mathematical abstractions to set up search in Lean for this
- (Roughly) Gradient-free methods for learning, such as Hebbian learning and in-context learning
- Learning Lean more rigorously
I’m reading Hitchhiker’s Guide to the Galaxy, The Handbook by Epictetus and Reinforcement Learning by Sutton and Barto.