Portrait of Xiaohan Yan

Xiaohan Yan颜小涵

Algorithm Engineer / Large Model @ AgiBot

I am an algorithm engineer working on large models at AgiBot, focusing on vision-language-action models for real-world robotic manipulation. Before that I received my M.S. from Tongji University, where I worked on 3D scene understanding and zero-shot instance segmentation. My interests are embodied AI, VLA / VLM and reinforcement learning — feel free to drop me an email if we share any.

Education & Experience

Experience
Algorithm Engineer / Large Model
2026 – Present
Algorithm Engineer / Large Model
2024 – 2025
Research Intern
Dec 2023 – Apr 2024
Education
M.S. in Computer Science
2022 – 2025
B.S. in Computer Science
Hohai University · ACM Team Captain
2018 – 2022

Publications

* indicates equal contribution

2026
VINE sampling overview
VINE: Taming Generative Control Policies for Reinforcement Learning
Rushuai Yang, Zhuo Han, Houlin Li, Hecheng Wang, Zhichao Wu, Rui Zhang, Zhaowei Zhang, Zihong Chen, Xiaohan Yan, Chiming Liu, Yi Chen, Wei Shan, Maoqing Yao
arXiv 2026 arXiv Project
An RL-oriented sampling method that re-anchors the interpolation state at every denoising step, enabling stable end-to-end value-gradient optimization of flow-matching policies.
ALOE framework overview
ALOE: Action-Level Off-Policy Evaluation for Vision-Language-Action Model Post-Training
Rushuai Yang*, Hecheng Wang*, Chiming Liu*, Xiaohan Yan, Yunlong Wang, Xuan Du, Shuoyu Yue, Yongcheng Liu, Chuheng Zhang, Lizhe Qi, Yi Chen, Wei Shan, Maoqing Yao
arXiv 2026 arXiv Project
Action-level off-policy evaluation for VLA post-training, enabling fine-grained credit assignment and stable policy improvement in real-world robotic manipulation.
HOLO pipeline overview
HOLO: Holistic Lightweight Optimization for Scene Understanding with Auto-Annotation and Multimodal Learning
Xiaoyun Hu*, Xiaohan Yan*, Nan Wang, Xiaowei Song, Gang Wei, Zhicheng Wang
WACV 2026
A large-scale scene description dataset together with a lightweight 3D-LLM.

Projects

LLM Science Exam competition
LLM Science Exam — Use LLMs to Answer Difficult Science Questions
Xiaohan Yan, Nan Wang, Xiaowei Song, Jinyu He
Kaggle 2023 Code
Fine-tuning large language models on private datasets. Score 0.905, top 3% worldwide, Silver Medal.
eScape game screenshot
eScape — A Geometry Storm Game
Origami-hui, Xiaohan Yan
GameJam 2023 Play
Scale your device and escape from this geometry storm. 1st in innovation and 2nd in theme interpretation.

Awards

Most of the awards I won during my student years 2018 – 2024
The 2019 ICPC Asia-East Continent Final — Bronze Medal 2019 – 2020
Jiangsu Collegiate Programming Contest — Silver Medal, 2nd place 2019 – 2020

News

Jul 2026 Our paper VINE is released on arXiv.
Feb 2026 Our paper ALOE is released on arXiv.
Sep 2025 Our paper HOLO is accepted by WACV 2026.
Mar 2025 I graduated from Tongji University with a Master's degree in Computer Science.

Misc

3D Modeling

Stuff I modeled and rendered in Blender.

Sports

Swimming · Flying disc

Games

Pokémon · Digimon

Language

Japanese