Patrick Jiang

Hello! I’m Patrick Jiang, a Ph.D. candidate in Computer Science at UIUC and a member of DMG. I’m advised by Prof. Jiawei Han and co-advised by Prof. Jimeng Sun.

I’m a Research Fellow at Anthropic, and was a Student Researcher at Google Research. I have also worked with Chroma, Google DeepMind, TML, and other industry partners.

My research centers on how language models reason and act as agents: how they seek information, use tools, and reason over external evidence. My prior work helped pioneer training search agents with on-policy reinforcement learning. More recently, I study the internal representations and cognition of modern language models, and how these insights can make them more reliable, controllable, and aligned.