About Me

Hi! I’m a second-year ELLIS Ph.D. student in computer science at Mila and Université de Montréal. I’m fortunate to be advised by the fantastic duo of Prof. Simon Lacoste-Julien and Prof. Moritz Hardt (at the Max Planck Institute for Intelligent Systems in Tübingen, Germany).

I’m broadly interested in understanding the societal impacts of AI from a theoretical perspective, using tools from learning theory, optimization, game theory and mechanism design. These days, I’m working closely with Prof. Gauthier Gidel on designing ranking algorithms for benchmarking and LLM evaluation based on pairwise preference evaluations. A key goal of this work is to make these rankings robust to strategic gaming, drawing on ideas from social choice theory.

Prior to grad school, I was a Pre-Doctoral Researcher at Google DeepMind India, where I worked with Dr. Rishi Saket and Dr. Aravindan Raghuveer. At Google, I worked on learning instance-level signals from aggregated feedback. This arises in Google Ads, where we may know how many people clicked an ad, but not who clicked it. We studied theoretically how to optimally aggregate data while preserving privacy, and also applied these ideas to derive fine-grained rewards for RLHF.

I spent four wonderful years as an undergraduate at the Indian Institute of Technology Delhi, graduating in 2023. I had the opportunity to spend my third-year internship and final year working at the Vector Institute and the University of Toronto, under the supervision of Prof. Rahul G. Krishnan. We developed methods for learning predictive checklists which have applications in clinical decision making.