Jiarui Yang

Jiarui Yang is a researcher affiliated with the Hong Kong University of Science and Technology (Guangzhou) and AGIBOT, where he works on vision-language-action models for robotic manipulation. He is a co-author of V-Link, a representation-recovery framework that improves Action DiT's access to 3D geometric and 2D semantic information from VLM features, demonstrating significant performance gains across simulation benchmarks and real-world humanoid tasks. His research also encompasses related work including multi-view camera field representations for VLA policies and reinforcement learning frameworks for long-horizon robotic manipulation.