PhD student, Department of Computer Science, University of Washington
1 paper at NeurIPS 2025
ReAgent-V enables reward-driven, multi-agent video understanding with dynamic reflection and frame selection.