Hi, I am an undergrad at the University of Washington, advised by Prof. Ranjay Krishna, PhD student Jieyu Zhang, and PhD student Zixian Ma at UW CSE RAIVN Lab. Currently I am also a student collaborator at Allen Institute for AI working on VLM training and data generation.

Research Interests:

• 2D and 3D Grounding and Tracking: Developing robust object detection, grounding and tracking systems that can understand spatial relationships and temporal dynamics across different modalities, with applications in autonomous systems and embodied AI.

• Compositional and Scalable Synthetic Visual Data Generation: Creating programmable and scalable synthetic data generation pipelines for computer vision tasks, focusing on scene graph programming and domain-specific data augmentation to improve model performance and generalization.

• Vision and Spatial centric VLM: Advancing vision-language models to better understand complex spatial reasoning, multi-step visual tasks, and tool-use scenarios, bridging the gap between visual perception and actionable intelligence.

📝 Selected Publications

(* denotes equal contribution)

Peer-reviewed

SOS: Synthetic Object Segments Improve Detection, Segmentation, and Grounding
Weikai Huang, Jieyu Zhang, Taoyang Jia, Chenhao Zheng, Ziqi Gao, Jae Sung Park, Ranjay Krishna
ILR+G @ ICCV 2025

Generate Any Scene: Evaluating and Improving Text-to-Vision Generation with Scene Graph Programming
Ziqi Gao*, Weikai Huang*, Jieyu Zhang, Aniruddha Kembhavi, Ranjay Krishna
SynData4CV @ CVPR 2025

ProVision: Programmatically Scaling Vision-centric Instruction Data for Multimodal Language Models
Jieyu Zhang, Le Xue, Linxin Song, Jun Wang, Weikai Huang, Manli Shu, An Yan, Zixian Ma, Juan Carlos Niebles, Silvio Savarese, Caiming Xiong, Zeyuan Chen, Ranjay Krishna, Ran Xu.
Blog at Salesforce Research | VentureBeat | MarkTechPost
SynthData @ ICLR 2025

Task Me Anything
Jieyu Zhang, Weikai Huang*, Zixian Ma*, Oscar Michel, Dong He, Tanmay Gupta, Wei-Chiu Ma, Ali Farhadi, Aniruddha Kembhavi, Ranjay Krishna.
NeurIPS 2024 (Main Conference) | Video-Language Models @ NeurIPS 2024 Oral Presentation
Blog at Snorkel AI | Talk at Snorkel AI

m&m's: A Benchmark to Evaluate Tool-Use for multi-step multi-modal Tasks
Zixian Ma, Weikai Huang, Jieyu Zhang, Tanmay Gupta, Ranjay Krishna
ECCV 2024

🎓 Education

• University of Washington
  B.S. in Computer Science, Minor in Applied Mathematics and Music
  2023 – 2027 (Expected)

📰 News

• [2024-9] Our work Task Me Anything is accepted to NeurIPS 2024.
• [2024-7] Our work m&m’s: A Benchmark to Evaluate Tool-Use for multi-step multi-modal Tasks is accepted to ECCV 2024.

🏅 Awards

• UW CSE John and JoAnne Wisniewski Endowed Scholarship 2024

📍 Professional Services

• Organizer of the Synthetic Data for Computer Vision Workshop at CVPR 2024 and CVPR 2025 (website)
• TA for UW CSE 455: Computer Vision, Autumn 2025 (course page)
• TA for UW CSE 493G: Deep Learning, Spring 2025 (course page)

📫 Contact

Email: weikaih@cs.washington.edu
Twitter (X): @weikaih04
Instagram: @weikaih04
Linkedin: Weikai Huang
Wechat: hwk18105962347