
Hi, I'm Chinmaya
I'm a researcher + engineer at Capital One.
I'm interested in building AI systems that are safe and useful by studying the abilities and limitations of language models. Currently, I'm a software engineer in the Applied Research group at Capital One, where I work on making language models better assistants through simulation, synthetic data, and evaluation. Previously, I worked on LM bootstrapping and self-improvement in Stanford's Computation & Cognition Lab, advised by Noah Goodman and Jan-Philipp Fränken.
I recently received my B.S. in Mathematical and Computational Science from
Stanford University, where I'm currently working on my M.S. in Computer Science (on leave).
Featured research publications
Latest research.
Featured projects
Some academic course projects, mini-investigations and miscellaneous interests.
Fixed PyTorch errors and byte-level bugs in Rust to allow REST to support modern LMs + large tokenizers (e.g. Llama 3).
2024
Investigating language models' ability to patch buggy code with the help of print statements.
2024