Jun Zhang (张峻)

I am a Researcher/Engineer at Joy Future Academy (京东探索研究院) through the TGT (Tech Genius Team) Program, working with Qingyi Si and Jiaqi Wang.

I am currently part of the LLM Foundation Model Team, focusing on:

  • Multimodal LLMs
  • Video Understanding (Video LLMs)
  • Post-training (RL & SFT)

I received my B.S. and M.S. degrees in Computer Science from Nanjing University in 2023 and 2026, respectively. I was a member of the MCG Group, advised by Prof. Limin Wang.

Previously, I was an intern in the Foundation Model Team at StepFun, and before that, a research intern at Tencent ARC Lab.

Open positions We are currently hiring for both full-time and internship positions for the TGT program. Please feel free to reach out if you are interested!

Portrait of Jun Zhang

News

Our new paper VideoChat3, a fully open video MLLM for efficient and generalist video understanding, is now on arXiv!

Check our new paper TimeLens. We expose critical quality issues in existing VTG benchmarks, propose quality-assured datasets, and release state-of-the-art open-source grounding models. Website ↗

p-MoD is accepted by ICCV 2025!

Publications

arXiv 2026

VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding

Xinhao Li, Yuhan Zhu, Xiangyu Zeng, Yuhao Dong, Haoning Wu, …, Jun Zhang, …, Yu Qiao, Yali Wang, Ziwei Liu, Kai Chen, and Limin Wang

StepFun 2026

Step 3.7 Flash: A High-Efficiency Flash Model for Real-World Agents

StepFun Team

CVPR 2026

TimeLens: Rethinking Video Temporal Grounding with Multimodal LLMs

Jun Zhang, Teng Wang, Yuying Ge, Yixiao Ge, Xinhao Li, Ying Shan, and Limin Wang

ICCV 2025

p-MoD: Building Mixture-of-Depths MLLMs via Progressive Ratio Decay

Jun Zhang, Desen Meng, Ji Qi, Zhenpeng Huang, Tao Wu, and Limin Wang

arXiv 2025

VideoCap-R1: Enhancing MLLMs for Video Captioning via Structured Thinking

Desen Meng, Rui Huang, Zhilin Dai, Xinhao Li, Yifan Xu, Jun Zhang, Zhenpeng Huang, Meng Zhang, Lingshu Zhang, Yi Liu, and Limin Wang

Experience

— Present

Full-time · Foundation Model Team at Joy Future Academy

Working on multimodal large language models and video understanding with Qingyi Si and Jiaqi Wang.

Intern · Foundation Model Team at StepFun

Worked on multimodal post-training and video understanding with Liang Zhao and Zejia Weng.