I am currently a Senior Research Scientist on the NVIDIA Cosmos Team. Previously I worked as a Senior Research Scientist at Meshy AI and Research Team Lead at Picsart AI Research. I’ve obtained my Ph.D. in 2023 from IFP Group at UIUC, supervised by Prof.Humphrey Shi after 2020, and formerly supervised by Prof. Thomas Huang before 2020.

Research Area

• Generative AI in Computer Vision.
• Large-scale omni-generative model including Text-to-Image, Text-to-Video and Text-to-3D.
• Caption, data, evaluation, and SFT for Text-to-Image.
• Specialized domain improvement such as T2I and T2V text rendering quality.

News

• [2026.06] Glad contributed on releasing the Cosmos 3 omni-model family. On Artificital Analysis leaderboard, our Cosmos3-Super-Text2Image is Ranked #1 on open-sourced and #4 among all T2I models including closed-source models [Model, Post, Paper]
• [2025.11] Join NVIDIA Cosmos team as Researcher.
• [2025.03] One paper accepted by ICCV 2025.
• [2024.02] Four paper accepted by CVPR 2024.
• [2023.07] Two paper accepted by ICCV 2023.
• [2023.06] Defenced my Ph.D. Thesis.
• [2023.05] Check out our new work Prompt-Free Diffusion: Taking” Text” out of Text-to-Image Diffusion Models, demo on the HuggingFace.
• [2023.03] One paper accepted by CVPR 2023.
• [2022.11] Check out our new work Versatile Diffusion: Text, Images and Variations All in One Diffusion Model, demo on the HuggingFace.