Kangning Liu (刘康宁)
Research Scientist @ Adobe | MLLM & Vision Foundation Models
Kangning Liu Profile Picture

Hello! I am Kangning Liu, currently a Research Scientist at Adobe. I earned my Ph.D. at the NYU Center for Data Science, where I was advised by Prof. Carlos Fernandez Granda and Prof. Krzysztof J. Geras. Before that, I earned my M.Sc. in Data Science from ETH Zurich and my B.E. in Electrical Engineering from Tsinghua University.

Research Focus: I build large-scale vision and multimodal foundation models that bridge cutting-edge research with real-world industrial impact.

My current work focuses on advancing multimodal understanding and integrating powerful visual comprehension into generative models. These frameworks enable a wide spectrum of applications, from GenAI data creation and reward modeling to equipping intelligent agents with complex tool-calling capabilities. My research has been published at top-tier venues (e.g., NeurIPS, CVPR, ICCV, ICLR) and deployed in production features across Adobe Photoshop, Lightroom, Express, and Firefly, impacting millions of users worldwide.

Recent Updates

  • Feb 2026: Our paper "VGent" and "Sparse-LaViDa" were accepted to CVPR 2026!
  • Jan 2026: Our paper "LaViDa-O" was accepted to ICLR 2026!
  • Dec 2025: Released "selectionByPrompt" feature in Adobe Photoshop integrating into ChatGPT.
  • July 2025: Our paper "RAS" was accepted to ICCV 2025!

For more details, feel free to reach out via email at kangning.liu[at]nyu.edu (still active). You can also find me on Google Scholar and LinkedIn .

Research Experience

Publications

See also Google Scholar.

Teaching

Service

Miscellaneous

Visual Fragments

Pageviews