Xunjian Yin

I am a CS PhD student at Duke University, co-advised by Prof. Shuyan Zhou and Prof. Bhuwan Dhingra. My research asks how AI systems can keep improving themselves over long horizons: agents that rewrite their own logic, learn from experience, and stay reliable when their environment changes. I focus on self-referential agents and world models.

Previously, I was at Google DeepMind, Amazon AGI SF Lab, the UCSB NLP Group (with Prof. William Yang Wang), Microsoft Research, and Peking University (with Prof. Xiaojun Wan).

I'm happy to discuss research and collaboration.