Ruoshi Wen
Ruoshi Wen is a robotics researcher at ByteDance Seed in Beijing, China, where he contributes to the company's dexterous manipulation program spanning hardware, teleoperation, and vision-language-action models. He is a co-author of the ByteDexter teleoperation paper and the GR-Dexter technical report, which together describe a 20-DoF anthropomorphic hand and a flow-matching VLA for dexterous manipulation. Prior to ByteDance, he conducted postdoctoral research at the University of Edinburgh focused on collaborative bimanual manipulation and interaction control.
“SiMDex uses only ~1.49M mined samples (<5% of the pool) yet improves the overall success rate from 47.7% to 61.1%”
Source→“Our base model follows GR-Dexter [47], a π0-like flow-matching VLA for dexterous manipulation.”
Source→“ByteDance developed GR-Dexter (the base VLA model used in this paper) and the ByteDexter hand”
Source→“Author of both GR-Dexter (the base VLA) and the ByteDexter teleoperation paper [48]. This person is building both the hardware interface and the learning system”
Source→“ByteDance developed GR-Dexter (the base VLA model used in this paper) and the ByteDexter hand... This paper signals ByteDance is building a full dexterous manipulation stack — hardware, VLA model, and now data curation infrastructure.”
Source→AI-extracted from podcast / newsletter / paper summaries. May contain errors.