Keyu Tian is an AI researcher and incoming Ph.D. student with eight years of experience focused on multimodal perception, video generation, self-supervised learning, and reinforcement learning. Currently a master’s candidate at Peking University, Keyu has held research internships at ByteDance AI Lab, SenseTime, and a remote collaboration with Oxford’s Torr Vision Group, blending industry-scale experimentation with academic rigor. They specialize in open-ended generative and RL problems for video and multimodal AI, regularly moving ideas from papers to prototypes in production-adjacent research environments. Based in Haidian, Beijing, Keyu pairs deep technical breadth with practical impact—recent work at ByteDance’s VC team targeted scalable video-centric models. An underrated strength is their consistent cross-institutional collaboration experience, enabling fast iteration across research labs and applied teams.
7 years of coding experience
Bachelor Software Engineering, Bachelor Software Engineering at Beihang University
[ICLR'23 Spotlight🔥] The first successful BERT/MAE-style pretraining on any convolutional network; Pytorch impl. of "Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling"
Contributions:2 releases, 16 commits, 109 pushes in 2 months
bertconvnethierarchicalsparsevisual-recognition
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.