Gordon Qian is a Senior Research Scientist at Snap Inc. with hands-on experience in multimodal generation across images, video, and 3D. He is a member of the video generation team, where he works on pretraining and post-training for 5B to 30B VLM-diffusion models, with a current focus on self-forcing for real-time streamable video generation. Previously, he served as tech lead for personalized generation, building end-to-end systems spanning large-scale data curation, model post-training, model evaluation, and product deployment. He earned his Ph.D. in Computer Science from KAUST, advised by Prof. Bernard Ghanem. He has authored 20+ top-tier publications with 3,500+ citations, filed six patents, and shipped three research innovations to Snapchat products, including Dreams, EasyLens, and AI Lens, impacting 900M+ monthly active users. His representative works include PointNeXt (NeurIPS, 1.2K+ citations, 1K+ GitHub stars), Magic123 (ICLR, 450+ citations, 1.6K+ GitHub stars), and Omni-ID (CVPR, product integrated, patent filed). His research interests include streamable and unified multimodal understanding and generation. He serves as an Area Chair for ICLR 2026.
If you are interested in working on image and video generative models with me, please reach out at gordonqian2017 [at] gmail.com