April 21, 2026
The Pierson Lab, within the Berkeley EECS department, is affiliated with the Berkeley AI Research Lab, Computational Precision Health, and the Center for Human-Compatible AI. See our lab website for more information on our group and recent work.
We seek applicants with methodological experience in language modeling or interpretability, and interests in applications in the health sciences, social sciences, or safe/reliable ML. Specific potential projects include:
1. Accelerating scientific discovery with LLMs and sparse autoencoders. This project will build on our work (ICML 2025, ICLR 2026) developing sparse autoencoder methods for hypothesis generation; we are pursuing several high-impact applications in the health and social sciences, in collaboration with clinicians and social scientists.
2. Using LLMs to provide safe and reliable health information. We will develop and deploy language models that can increase access to health information (New England Journal of Medicine AI 2025, NeurIPS 2024).
Applicants will collaborate closely with PI Emma Pierson, with other PhD students and postdocs in the lab, and with clinicians, social scientists, and other computer science faculty.