I’m a graduate student at UC San Diego, where I did my bachelors too. I’m advised by Professor Zhuowen Tu and Professor Zhiting Hu.

My main research interests are in spatial intelligence, world models, and interpretability. I’m particularly interested in scene reconstruction, and in building environments where an agent has to learn a latent model of the world's dynamics (often called a ‘world model’).

I believe understanding the causal structure encoded in these learned world models is crucial for building agents that can reason about interventions, generalize beyond their training environments, and move closer toward more general forms of intelligence.

I’m working on post-training vision language models at Typeface this summer.

I also worked as a machine learning engineer at Tesla, with the energy team.