My current work spans from theoretical foundations in AI to practical applications in AI-assisted science. On the theoretical side, I study fundamental phenomena in deep learning such as emergence and scaling laws in neural networks (e.g., see [here] for our work on an exactly solvable model for emergence and scaling laws). On the applications side, I'm developing AI systems for scientific reasoning, including agents that can debate experimental strategies in particle physics, a framework designed to generalize across scientific domains (explore the project on [GitHub]).

​

Previously, I worked at the intersection of particle physics and cosmology, exploring how the early universe informs our understanding of fundamental physics. See [here] for all publications.