Xianjin Wu

portrait.jpg

I am at Huazhong University of Science and Technology (HUST). During my undergraduate studies, I was fortunate to be advised by Xiang Bai and Dingkang Liang.

My research interests lie in 3D scene understanding, spatial intelligence, and embodied AI, with a particular interest in machine learning systems (MLSys). I am interested in building intelligent systems that can understand the physical world and act efficiently within it.

My recent work spans spatial understanding with VEGA-3D, real-time robotic manipulation with TurboVLA, and training-free policy acceleration with RePIC. I also work on multi-camera 3D perception and inference optimization through research and competitions.

I am currently a research intern at DiDi Voyager, and previously interned at Baidu.

Email · CV · Google Scholar · GitHub · ORCID

News

Nov 2026 Global Runner-Up (2/1,855 teams) — Team EVA won RMB 60,000 at the inaugural Robbyant Embodied AI Challenge on November 16, 2026. Tencent Technology (腾讯科技), Xinzhiyuan (新智元), and QbitAI (量子位) interviewed the team for their WeChat coverage. Read on WeChat. Robbyant (蚂蚁灵波) AgileX Robotics (松灵机器人)
Aug 2026 Our team won 1st place in the AI City Challenge 2026, Track 1! GaugeAlign is available with code and our technical solution.
Jul 2026 We released TurboVLA, a compact vision-language-action model for real-time robotic manipulation.
Jun 2026 VEGA-3D was accepted to ECCV 2026!

Selected Publications

  1. vega.png
    Generation Models Know Space: Unleashing Implicit 3D Priors for Scene Understanding
    Xianjin Wu, Dingkang Liang, Tianrui Feng, Kui Xia, Yumeng Zhang, Xiaofan Li, Xiao Tan, and Xiang Bai
    In European Conference on Computer Vision (ECCV), 2026
  2. repic.png
    Don’t Throw Away the Solution: Reusing Denoising Trajectories for Fast Robot Control
    Xianjin Wu, Xian Li, Jianxiang Tang, Dingkang Liang, and Xiang Bai
    arXiv version forthcoming, 2026
  3. turbovla.png
    TurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM
    Hengyi Xie*, Chenfei Yao*, Xianjin Wu, Yingying Zhu, Dingkang Liang, Xiang Bai, and Han Ding
    arXiv, 2026
  4. gaugealign.png
    The Winning Solution for Track 1 of the 2026 AI City Challenge: Gauge Canonicalization for Multi-Camera 3D Perception
    Xianjin Wu*, Heng Fang*, Xuanyang Xi, Yiping Tang, Di Xu, Xiang Bai, and Dingkang Liang
    In ECCV Workshop: 10th AI City Challenge, Oral, 2026

Honors & awards

Kaggle Expert (3,590 of 225,021) · HyperbolicCurve

2026

2025

Industrial Experience

DiDi Voyager Research Intern

August 2026 – Present

Mentors: Xuesong Chen and Shaoshuai Shi.

Baidu Research Intern

August 2025 – April 2026

Mentors: Kui Xia and Yumeng Zhang.

Education

Huazhong University of Science and Technology

B.Eng. in Artificial Intelligence

Sep 2022 - Jun 2026

Wuhan, China

Academic service

Reviewer: PR, ECCV, ICLR.