Jianwen JIANG
jianwen.alan [at] gmail.com
jianwen.alan [at] gmail.com
I am working on interactive multimodal intelligence at Meta Superintelligence Lab. Before this, I led the human-centric generation efforts at ByteDance, and earlier worked on perception at Alibaba.
Before joining Meta, my earlier work at Alibaba on video understanding took first place in six CVPR challenges (recognition, temporal localization, and spatio-temporal detection); the resulting models were deployed in Alibaba Cloud and used at the Beijing Winter Olympics. At ByteDance, I led Loopy and the OmniHuman series, and contributed to Seaweed and Seedance, shipped to billions of users.
Projects I led have been covered by Forbes, ABC, Wired, Business Insider, VentureBeat, CNR News, Synced, and others, and featured as the “first preview” of this new generation in a16z's “AI Avatars Escape the Uncanny Valley”. Currently, I'm passionate about an AI-native working style, enjoy small teams that operate like startups.