Prethiv Sriman

I'm a junior at SRMIST working in AI research.

— Primarily interested in Deep Learning, Mechanistic Interpretability, and Physics-Informed Machine Learning

I'm particularly interested in understanding what neural networks learn beyond what we explicitly teach them.

— Using AI to uncover hidden variables and structure in high-dimensional systems

— Understanding representations that are difficult for humans to perceive directly

I co-created Ratiod, which is a cool looking academia wrapper with 10k+ active researchers, 250k+ papers indexed, and 99.4% citation extraction accuracy.

In my lifetime I want to:

— Solve AGI

— Make dinosaurs

— Build a teleportation device

— Make virtual worlds for humans

— Live on another planet

— Meet 100,000+ unique individuals

— Live for 1000+ years

“I'd like to work towards a view of AI not just as a tool for prediction, but as a way of discovering structure that we otherwise cannot see. also i hate working with agentic ai stuff, but i still do T-T”