Summary
Xiangtai Li is a Staff Research Scientist based in Singapore with nine years of experience building computer vision, deep learning, and multi-modal models across industry and academia. He earned a Ph.D. from Peking University and has held research roles at ByteDance (TikTok), NTU MMLab, SenseTime, JD, and Shanghai AI Lab, contributing to visual segmentation, detection, open-vocabulary learning, and multi-modal LLMs. At TikTok he has driven TT-Live business model development and led work on Pixel-MLLM and single-transformer architectures for vision and language, while open-sourcing much of his research on GitHub. Known for blending strong engineering instincts with research rigor, he mentors students and builds dataset and infra tooling to move models toward production. He is actively open to remote collaborations and emphasizes efficient, scalable universal models as his primary passion.
9 years of coding experience
2 years of employment as a software developer
学士, 物联网工程, 学士, 物联网工程 at Beijing University of Posts and Telecommunications
博士, 计算机科学与技术, 博士, 计算机科学与技术 at Peking University
学士, 物联网工程, 学士, 物联网工程 at Queen Mary University of London