Hello! I'm Vivek. I research, build, and evaluate LLMs. I write on Reinforcement Learning, LLMs, and engineering.

I work at CoreWeave, where I research RL techniques and build open-source libraries for training LLM agents.

Previously, I co-founded Cartograph, building LLM agents for codebase understanding, and led the AI team at Handshakes, where I built systems that extracted insights from unstructured data to power search and monitoring. I worked on recommendation systems, information extraction, knowledge graphs, search, and LLMs.

You can email me at hello@ this domain.