Building real world AGI.
Currently, I am Head of Pretraining at AMI Labs.
I was a research lead at OpenAI, where I led research on video perception. I was a core contributor to several of OpenAI's flagship models: DALL-E 3, Sora 1, GPT-4o, Thinking with Images, 4o Image Generation, and GPT-5.
I earned my PhD in physics from MIT.
Previously, I cofounded Lightelligence (Xizhi Tech), now listed on HKEX.
I've been recognized as MIT TR35, Forbes 30 Under 30, and won a Gold Medal at IPhO.