About

I am an incoming Master’s student at Peking University, advised by Prof. Daquan Zhou. I am currently a member of DAGroup.

Previously, I worked on LLM post-training with Prof. Ning Ding at Tsinghua University, audio-visual generation and editing with Prof. Di Hu at Renmin University of China, and the economics of generative models with Prof. Tonghan Wang at Tsinghua University.

My journey in computer science began with competitive programming, which gave me a strong algorithmic foundation that has naturally carried over into my research. My current research focuses on agents and robotics.

I am always happy to meet new friends and collaborators. Feel free to reach out via email at ruiyangsi0329@gmail.com or add me on WeChat.

News

  • 2026-10 Released PyRUA-Lean, an efficient code-execution framework for robot-use agents: higher success with far fewer tokens, state-of-the-art on RoboDojo, and a trained 4B Coder model that generalizes to unseen tasks. [paper] [project] [code]
  • 2026-06 Received the SenseTime Scholarship, awarded to up to 30 undergraduates nationwide in AI.

Selected Publications View all →

* Equal contribution   † Corresponding author

PyRUA-Lean

Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success but 65% Fewer Tokens

Ruiyang Si, Jianxin Bi, Shunyu Yang, Rui Ni, Wenbo Huang, Qiang Wang, Shulong Jiang, Yintianrun Zhang, Feixiang Ruan, Kewei Zhang, Duomin Wang, Xiuyu Li, Haiwen Feng, Zhen Dong, Daquan Zhou†

Technical Report | arXiv | Project | Code | Model

An efficient code-execution framework for robot-use agents: the agent chains robot primitives in Python cells, raising success from 63.1% to 71.7% with 65% fewer input tokens. Achieves state-of-the-art on RoboDojo (51.6% success, 19.4 points above the previous best on the public leaderboard), and our GRPO-trained 4B Coder generalizes to unseen tasks (18.8% → 44.1%), outperforming Qwen3.5-27B.

Education

Peking University Incoming
Master's student, advised by Prof. Daquan Zhou
Beijing University of Posts and Telecommunications 2023 – 2027 (expected)
B.S. in Computer Science and Technology · Rank 1/376

Honors & Awards

Scholarships

  • 2026.10National Scholarship, Ministry of Education of China
  • 2026.06SenseTime Scholarship (up to 30 undergraduates nationwide in AI)
  • 2025.10CCF Elite Collegiate Award, China Computer Federation (99 undergraduates nationwide)
  • 2025.10National Scholarship, Ministry of Education of China
  • 2024.10National Scholarship, Ministry of Education of China

Competitions

  • 2024.10Gold Medal, The 2024 ICPC Asia Regional Contest
  • 2024.10Gold Medal, The 10th China Collegiate Programming Contest (CCPC)
  • 2024.10Gold Medal, CCF Collegiate Computer Systems & Programming Contest (CCSP)
  • 2021.07Silver Medal, The 38th National Olympiad in Informatics (NOI 2021)