hey

I'm Sriraam, an Applied Researcher on the post-training team at Chakra Labs. I build RL environments, training tasks, and internal benchmarks focusing on knowledge work for foundation labs.

These are my research interests:

• Sample efficiency
• Continual learning
• Human simulation
• Mech interp + RL
• World models and neural RL envs
• Sycophancy and adaptive reasoning

If you find any of these topics interesting I'd love to talk to you.

My time studying Learning Science at Harvard has made me see almost every problem as a learning problem. I draw inspiration from human learning for my Reinforcement Learning work.

I like Hogwarts, Anime, and K-dramas.