Su Jiayi | 苏佳奕

Hiii, there!

I am currently a first-year Ph.D. student at CASIA, Galbot, and EPIC Lab (2026-present). Before that, I received my B.Sc. degree in Computer Science and Technology from Xiamen University (2022-2026). I am currently an intern at Galbot, under the supervision of Prof. He Wang and Prof. Zhizheng Zhang.

I am interested in embodied AI, VLA for manipulation, world models, and video models for robotics.

I am lucky to work closely with Mi Yan, Jiangran Lyu, and Shengliang Deng.

Email / Github / X / OI Blog / Luogu

Publications

*: joint first author; † project lead; ✉ corresponding author(s)

	GPT 6 Astra as an Embodied Policy Jiayi Su^, Yixin Zheng^, Mi Yan, Li Yi, Zhizheng Zhang^✉, He Wang^✉ Technical Report, 2026 project page / code / 具身智能之心 / 腾讯科技 / 智能纪元AGI / human five A technical report comparing direct GPT 6 Astra control with a hybrid π0.5 + GPT 6 Astra policy for zero-shot bimanual manipulation. On the selected RoboDojo tasks, the hybrid policy reaches 48% success and a 62.60 mean score, compared with 26% and 37.81 for direct control.
	ZETA: A Controlled Study of Zero-Shot Cross-Embodiment VLA Transfer for Tabletop Manipulation Mi Yan^, Wenhao Zhang^, Zhiqi Zhang^, Yu Peng^, Tangxinyu Wang^, Lingfei Zhai, Jiayi Su, Shengliang Deng, Lin Peng, Yaowei Liu, Yuxing Chen, Zhiyuan Wei, Jilong Wang, Jiayi Chen, Jiangran Lyu, Zhizheng Zhang^✉, He Wang^✉ CoRL*, 2026 project page / paper / code ZETA presents a controlled study of zero-shot cross-embodiment transfer for vision-language-action models in tabletop manipulation.
	StereoVLA: Enhancing Vision-Language-Action Models with Stereo Vision Shengliang Deng^, Mi Yan^, Yixin Zheng^, Jiayi Su, Wenhao Zhang, Xiaoguang Zhao, Heming Cui, Zhizheng Zhang^✉, He Wang^✉ RSS*, 2026 arXiv / code / checkpoint StereoVLA is powered by stereo vision and supports zero-shot deployment with high tolerance to camera pose variations.
	ArtFormer: Controllable Generation of Diverse 3D Articulated Objects Jiayi Su^{, ✉}, Youhe Feng^, Zheng Li, Jinhua Song, Yangfan He, Botao Ren, Botian Xu^✉ CVPR (top 15%), 2025 arXiv / code ArtFormer introduces a novel transformer-based framework that generates diverse, high-quality 3D articulated objects from text description or single image.

Experience

	University of Chinese Academy of Sciences Ph.D. Student at CASIA (Institute of Automation) 2026.09 - present
	Galbot Large Embodied Model Researcher 2025.08 - present VLA & Simulation Pipe Engineer
	Shanghai Jiao Tong University Research Assistant 2025.02 - 2025.05 Research focus: DiT inference optimization
	Xiamen University B.Sc. in Computer Science and Technology 2022.09 - 2026 Rank: 1/71; Grade: 97.0 / 100 (3.88 / 4.00)

Service

CVPR 2026 Reviewer
ICME 2025 Reviewer

Awards & Honors

2022-2025 First-Class Scholarship (Xiamen University)
2021 National Olympiad in Informatics (NOI), Winter Camp, Silver Medal
2021 National Olympiad in Informatics (NOI), Bronze Medal
2020-2022 National Olympiad in Informatics in Provinces (NOIP) First Prize, Rank 10

Style adapted from Jon Barron.