hi, i'm prashant!

I’m interested in how we can represent the physical world so models can reason about it and act more reliably in the real world.

Unpopular opinion – LLMs don't reason and are at the end of the day just next token predictors. Reasoning is philosophical and goes beyond mere definition, read more here.

I'm currently seeking research-oriented roles both in academia and industry around physical AI, reasoning models, generative models and GPU acceleration.