Hey, I'm Rushikesh

Most of what I build today revolves around one question: How should AI model people?

That means building systems that construct rich, evolving models of people from behavioral signals, causal chains, and lived interactions, bringing the idea of world models to human understanding.

But I'm increasingly interested in the layer underneath personalization: the internal representations an AI needs of the people, and eventually the entities, it is reasoning about. That's where I think ideas from world models, memory, and evolving ontologies start to converge.

I didn't start there, though. I'm a CS undergrad who's spent a lot of time wrestling with hard problems until they become products. Along the way I've built a vaultless password manager that eliminates database exfiltration risks, a career intelligence platform that produces evidence backed AI-powered career insights, and an RL hiring environment that exposes long-horizon AI reasoning gaps in frontier models.

Today, I'm building an AI-native personal intelligence product based on this premise. It has pulled me deep into AI memory architectures, temporal knowledge graphs, and unsupervised ontologies.

Always happy to talk AI, systems, startups, or ambitious ideas. Reach out on X, LinkedIn, or via email.