Skip to content

Sudarsh

I work on reasoning systems, inference optimization, and multimodal AI. My world model is simple: build systems that expose how models think, not just what they output.

Selected work

Recent writing