I'm a physics PhD student at UC Berkeley studying capability expression in AI systems. My research currently focuses on understanding how language models acquire new capabilities and adapt to new tasks, with a particular emphasis on the information theoretic requirements for elicitation of latent capabilities vs. teaching new skills and the impact of different training paradigms (e.g. pretraining, RLHF) on capability acquisition. I'm interested in how these insights can be applied to improve the performance and safety of AI systems, particularly in the context of large language models.

Prior to this, I developed and applied novel optical techniques to study emergent phenomena and exotic phases of matter in quantum materials. My research focused on how light and material systems interact when they are affected by polarization and magnetic field, and I used various optical techniques at the visible and near-infrared regions to study emergent phenomena in strongly correlated electron systems.

In my free time, I enjoy running, film photography, music production, analog video synthesis, and collecting other creative hobbies.