(Note: I am applying to PhD programs this cycle, please reach out to me at hrithikr@umich.edu!)

Hi, nice to meet you! I am broadly interested in developing prescriptive theories for language modeling. I aim to gain an actionable understanding of deep learning and LLMs using mathematical and empirical techniques, thus enabling principled algorithmic innovation. Currently I'm working with Ambuj Tewari, analyzing different ways to elicit chain-of-thought.

I finished my Bachelor's and Master's at the University of Michigan, advised by Clayton Scott and Rada Mihalcea. During this time, I explored advanced coursework in theoretical machine learning and LLMs, and published a first-author NeurIPS 2024 paper on implicit bias. Recently I wrote a benchmarking research proposal accepted to Google DeepMind (GSoC 2025), but I dropped out of the program to focus on research more aligned with my interests.

Please see my CV for other highlights!

1