Weikai Huang
PhD Student @ UW CSE, RAIVN Lab
I am a PhD student at 
UW, working on Computer Vision and Robotics.
I'm advised by Prof. Ranjay Krishna, and have been fortunate to work closely with Prof. Jason Ren, Prof. Bharath Hariharan, Prof. Ali Farhadi, Dr. Jieyu Zhang, and PhD student Zixian Ma. I am also currently a research intern at 
Salesforce AI Research. Before the PhD I did my B.S. at UW and was a student researcher at 
AI2.
I'll be at CoRL 2026 (Austin, Nov 9–12) and NeurIPS 2026 (Atlanta, Dec 9–13) — feel free to reach out for a coffee chat! I'm also happy to connect with students interested in 3D vision, embodied AI, generation, and unified models. Reach out.
News
- Oct 2026Reflex is out: visually grounded reactive humanoid control for catching human-thrown boxes.
- Sep 2026MolmoAct2 accepted to CoRL 2026 as a Spotlight. See you in Austin!
- Jun 2026Two papers, IPT and Synthetic Visual Genome 2, accepted to ECCV 2026.
- Jun 2026Joining Salesforce AI Research as a Research Intern in Palo Alto, June–September 2026.
- Jun 2026Molmo2 received the CVPR 2026 Compute Transparency Champion award.
- May 2026MolmoAct2 released: action reasoning models for real-world robot deployment, with code, models, and data.
- Apr 2026WildDet3D released: promptable 3D detection in the wild. Try the demo or the iPhone app.
- Apr 2026Molmo2 selected as a Best Paper Candidate at CVPR 2026.
- Feb 2026Three papers, SOC, Molmo2, and TrajTok, accepted to CVPR 2026.
- Jan 2026Generate Any Scene accepted to ICLR 2026.
- Sep 2024Task Me Anything accepted to NeurIPS 2024.
- Jul 2024m&m's accepted to ECCV 2024.
Publications * equal contribution · § core contributor
Experience
- Salesforce AI ResearchResearch Intern · with Ran Xu, Zeyuan Chen, Silvio Savarese, An Yan, Luyu Yang3D Generation & Editing, Streaming Video Understanding, and Interactive Models2026–
- Allen Institute for AIStudent Researcher · with Ranjay Krishna, Ali Farhadi, Dieter Fox, Jason Ren, Christopher Clark, Sangho LeeOpen-Source VLMs, VLA Post-Training, VLM Training Infra, 3D Perception, and Data CrowdsourcingWildDet3D, MolmoAct2, Molmo22025–26
Education
- University of WashingtonPh.D. in Computer Science2026–
- University of WashingtonB.S. in Computer Science, with Honors, cum laude2023–26
Talks
- Foundation Models for RoboticsMathWorksOct 2026
- What GPT-6 Astra Means for Robotics—and What to Scale NextUnderstanding AISep 2026
- The Right Representation for 3D LearningSalesforce AI ResearchAug 2026
- Synthetic Grounding Data That Beats Real DataAI2Apr 2025
Awards
Teaching
- TA, CSE 455 Computer VisionAu 25
- TA, CSE 493G Deep LearningSp 25, Wi 26
Service
- Organizer, SynData4CV @ CVPR2024–26
- ReviewerCVPR, ECCV, NeurIPS, CoRL, AAAI, WACV, BMVC, RA-L, 3DV