Benjin ZHU

Ph.D

Chinese University of Hong Kong

Biography

Benjin Zhu is a Research Scientist at Li Auto and a postdoctoral researcher at Tsinghua University with Prof. Jifeng DAI. He received his Ph.D. from the Department of Electronic Engineering, CUHK in 2025, advised by Prof. Hongsheng LI and Prof. Xiaogang WANG in the MMLab, and his B.Eng. from SCUT in 2018. Before his Ph.D., he was at MEGVII Research with Dr. Gang Yu, Dr. Xiangyu Zhang, and Dr. Jian Sun.

His current research focuses on cross-embodiment VLA and World Models with RL. He led the team that won the inaugural nuScenes 3D Object Detection Challenge at WAD, CVPR 2019, and authored widely-used open-source frameworks Det3D, CVPods, and EFG.

Interests

Embodied AI
Physical Intelligence

Education

Ph.D in Electronic Engineering, 2021 ~ 2025
The Chinese University of Hong Kong (CUHK)
B.Eng in Software Engineering, 2014 ~ 2018
South China University of Technology (SCUT)

News

2026-05 Released the Mind-Omni series — our L4 autonomous driving stack covering VLA, World Models, and RL. ✨
2025-06 Graduated from CUHK MMLab and joined Li Auto as a Research Scientist on L4 autonomous driving.

Work Experience

Senior Research Engineer

Li Auto

May 2025 – Present Beijing, China

World Models, Vision-Language-Action Models, Reinforcement Learning.

Researcher

MEGVII Research

January 2019 – May 2021 Beijing, China

End-to-end Object Detection, Unsupervised/Self-supervised Learning, Research Infrastructure.

Publications

Publications list can be found HERE. Or visit my Google Scholar.

Yuzhou Huang, Benjin ZHU, Hengtong Lu, Victor Shea-Jay Huang, Haiming Zhang, Wei Chen, Jifeng Dai, Yan Xie, Hongsheng Li

May, 2026 arXiv

MindVLA-U1: VLA Beats VA with Unified Streaming Architecture for Autonomous Driving

A unified streaming VLA architecture that surpasses experienced human drivers on WOD-E2E for the first time, at VA-level latency.

Benjin ZHU, Zhe Wang, Hongsheng Li

July, 2024 In ECCV

nuCraft: Crafting High Resolution 3D Semantic Occupancy for Unified 3D Scene Understanding

ECCV 2024, Dataset

Benjin ZHU, Zhe Wang, Shaoshuai Shi, Hang Xu, Lanqing Hong, Hongsheng Li

December, 2022 In CVPR

ConQueR: Query Contrast Voxel-DETR for 3D Object Detection

CVPR 2023 Highlight Presentation (top 2.5%)

Benjin ZHU, Jianfeng Wang, Zhengkai Jiang, Fuhang Zong, Songtao Liu, Zeming Li, Jian Sun

July, 2020 arXiv

AutoAssign: Differentiable Label Assignment for Dense Object Detection

End-to-end dense object detection without hyperpamameters.

More Publications

Pengfei Jing, Victor Shea-Jay Huang, Hengtong Lu, Jifeng Dai, Xie Yan, Benjin ZHU (2026). Action Emergence from Streaming Intent. arXiv.

PDF Cite Project

Hengtong Lu, Victor Shea-Jay Huang, Chengmin Yang, Pengfei Jing, Jifeng Dai, Yan Xie, Benjin ZHU (2026). Driving Intents Amplify Planning-Oriented Reinforcement Learning. arXiv.

PDF Cite Project

Benjin ZHU, Xiaogang Wang, Hongsheng Li (2025). ConsistentCity: Semantic Flow-guided Occupancy DiT for Temporally Consistent Driving Scene Synthesis. In ICCV 2025.

Cite Code Project

Benjin ZHU, Junqiang Huang, Zeming Li, Xiangyu Zhang, Jian Sun (2020). EqCo: Equivalent Rules for Self-supervised Contrastive Learning. arXiv.

PDF Cite Code

Benjin ZHU, Zhengkai Jiang, Xiangxin Zhou, Zeming Li, Gang Yu (2019). Class-balanced Grouping and Sampling for Point Cloud 3D Object Detection. arXiv.

PDF Cite Code

Projects

For all projects, see here.

EFG: An Efficient, Flexible, and General deep learning framework that retains minimal.

Easy-to-use research codebase. Users can use EFG to explore any research topics following project templates.

CVPods: All-in-one Toolbox for Computer Vision Research.

Welcome to cvpods, a versatile and efficient codebase for many computer vision tasks. The aim of cvpods is to achieve efficient experiments management and smooth tasks-switching.