Hey there!

I'm Sree Sharvesh! Currently interested in AI Safety Research. My primary focus is on AI Control and Interpretability. Learn more about my work here..

My research focuses on

• Control Evaluations
• Multi-agent systems and alignment
• Few-shot catastrophe prevention
• Agentic evaluations, adversarial tests, and sandboxed tool-use environments
• LLM cognition: belief dynamics, CoT structure, and reasoning circuits

I dislike being bored! I keep doing something so that I don't get bored.

Outside work, I spend my time hunting for new cuisines, digging through music and watching anime that steals more hours than I plan. I still show up for hackathons, CTFs, and late night CP sessions when I miss the thrill of doing something just for fun.

Reach out if you'd like to talk projects, research or something interesting.

Yours,
Sree Sharvesh