Browse publications on Google Scholar (top-right) ↗
Resume (English only)
Academic Achievements
Published multiple papers in top conferences and journals such as ICML, ICLR, NeurIPS, TPAMI, CVPR, ECCV, and ACMMM. Notable works include FAST-VQA and DOVER, with OneAlign scorer downloaded over 600K times on HuggingFace. Involved in projects like Kimi-VL-Thinking-2506, Kimi-VL, and Aria.
Research Experience
Currently a researcher at Moonshot AI, leading the Kimi-VL team, developing open-source and proprietary models, and pursuing frontier basic visual abilities. Formerly the lead of the Q-Future project, focusing on visual evaluation with LMMs.
Education
B.S.: Peking University (Computer Science); Ph.D.: Nanyang Technological University (Singapore), supervised by Prof. Weisi Lin.
Background
Research interests include LMM pre-training, long-prefill, and long-decode extensions. Currently working at Moonshot AI with Xinyu Zhou. Previously a PhD candidate at Nanyang Technological University, supervised by Prof. Weisi Lin. Obtained B.S. in Computer Science from Peking University.
Miscellany
Interests include fine-grained visual perception, video understanding, multimodal reasoning, and multimodal long-context understanding.