I work on how language models decode: diffusion LMs at Conscious Engines, before that mixture-of-experts neural operators at IBM Research. Dual degree in EE at IIT Bombay.
- COSMOS: first-author paper, NeurIPS 2026 workshop (STODY). Soft mechanism mixtures for long-horizon PDE forecasting.
- PriorityKV: KV-cache compression that keeps an agent's tool calls intact. 0.688 vs 0.125 reliability at the same memory budget.
- Ghar: describe a room out loud, get a designed home back. Built on Gemini Live. Demo
- Sole engineer behind a hotel operating system running at live properties.
anupamrawat.com · X · LinkedIn · anupam.rawat.iitb@gmail.com