I am an Assistant Professor at KAIST AI leading the Computation & Cognition (COCO) Lab. Concurrently, I am a Research Manager at NVIDIA and will serve as the Head of the NVIDIA-KAIST Joint AI Research Lab. If you are interested in joining my lab, please visit this page.

My research takes a data-centric view of intelligence: what a model can learn is fundamentally shaped by the data it experiences. From this perspective, I study how we can create, select, and structure better learning experiences for machine cognition. A central theme of my work is social intelligence, because other intelligent systems provide uniquely rich and efficient sources of data, feedback, and experience.

This motivates two complementary directions in my research. (1) Machine social learning: I study how models can develop stronger cognitive capabilities by learning from other models and from themselves through synthetic data, distillation, self-improvement, and inference-time interaction. (2) Social world modeling: I study how machines can understand and reason about people and other agents—their beliefs, intentions, relationships, norms, and behaviors.

Before joining KAIST, I spent a gap year as a Postdoctoral Researcher working with Yejin Choi and Jan Kautz at NVIDIA. I was a Young Investigator and an intern at the Allen Institute for AI (Ai2), and received my Ph.D. in Computer Science from Seoul National University, advised by Gunhee Kim.