My name is Hari. I’m interested in how systems can help people live more meaningful lives. I am particularly focused on the future of embodied intelligence and materializing purposeful acceleration towards the future.

I currently work on frontier AI at OpenAI, with a focus on enabling models to follow human supervision and learn while working autonomously for long periods of time. I previously conducted research on sample efficient robotic learning, reinforcement learning, and meta learning through affiliations with Berkeley Artificial Intelligence Research and Google DeepMind, and currently work on novel pretraining architecture research to train models with physical intuition.

Outside of research, I’m a strategic angel investor backing exceptional people across technology, science, and frontier industry. I also hold formal and informal advisory roles spanning venture, policy, and culture.

Prior to university, I founded a company that was acquired when I was 17 and conducted internationally recognized research in RL, CV, and robotics. I’m an currently an intermittent student at UC Berkeley’s M.E.T. program. If you think we’d have an interesting conversation, feel free to reach out :)