I’m a PhD Candidate in Computer Science at Harvard University. I am interested in understanding when and why deep learning works well, and what we need to do to make it better. In particular, my PhD research has been focused on understanding generalization and representations by using a mix of theory and empirics.

I’m fortunate to be advised by David Cox and Boaz Barak. I am an organizer of the Harvard ML Theory Group. I have a strong side interest in computational neuroscience and computational cognitive science.

I received my bachelors and masters in Electrical Engineering at IIT Bombay.

Note: I am actively looking for full-time positions in industry and academia. Please get in touch!

News

• Our paper Revisiting Model Stitching to Compare Neural Representations has been accepted to NeurIPS 2021
• I have been awarded the Siebel Scholarship — awarded annually for academic excellence and demonstrated leadership to over 90 top students from the world’s leading graduate schools.
• I am interning with the Behrooz Ghorbani (Google Translate) and Behnam Neyshabur (Google Blueshift) this summer on scaling laws in neural machine translation.