I am a first-year PhD student in Computer Science at the Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University, supervised by Professor Wei Xu. I received my bachelor’s degree in Artificial Intelligence from Yuanpei College, Peking University, in 2026.

My research focuses on reinforcement learning and large language models (LLMs), with a particular interest in developing agentic algorithms and methods for LLM post-training. My research interests also include AI alignment and interpretability. My research is driven by the following question:

How can we close the gap between artificial intelligence and human-level intelligence by designing reliable AI systems and efficient learning algorithms?

Education

• Tsinghua University, Institute for Interdisciplinary Information Sciences (IIIS)
  PhD Student in Computer Science and Technology; advisor: Prof. Wei Xu
  September 2026 – Present
• Peking University, Yuanpei College / Tong Class
  B.S. in Artificial Intelligence
  September 2022 – July 2026

Research Experience

• Ant Research Reinforcement Learning Lab
  Research Intern; advisor: Prof. Yi Wu
  June 2025 – June 2026
  Research on reinforcement learning and LLM post-training.
• Tsinghua MARS Lab: Multimedia Computing, Autonomous Driving, Robotics and Sensors
  Visiting Student Researcher; advisor: Prof. Hang Zhao, IIIS, Tsinghua University
  February 2025 – June 2025
  Research on autonomous driving and multimodal models.
• PAIR Lab: PKU Alignment and Interaction Research Lab
  Visiting Student Researcher; advisor: Prof. Yaodong Yang, Peking University
  October 2023 – February 2025
  Research on alignment and reinforcement learning.

Publications

• (arXiv Preprint, 2026) Humanize MLE Agents: Multi-Agent Session Alternation Beyond Single-Agent Limits
  Changye Li, Zijian Zhang, Muheng Li, Yiduo Pang, Tingjun Wu, Menghan Li, Yitong Liu, Hongkai Guo, Mengchi Chen, Chenyu He, Kaisheng Ma, Wei Xu, Song Han, Sihao Liu, Ligeng Zhu
• (arXiv Preprint, 2026) Humanize: Judgement Engineering for Agentic Coding
  Sihao Liu, Ligeng Zhu, Zijian Zhang, Dongyun Zou, Zhengyang Zhang, Changye Li, Song Bian, Song Han, Tony Nowatzki
• (NeurIPS 2026 Poster) Perceive, Interact, Reason: Building Tool-Augmented Visual Agents for Spatial Reasoning
  Changye Li, Meng Lu, Yi Wu, Ligeng Zhu
• (ECCV 2026 Poster) EGM: Efficient Visual Grounding Language Models
  Guanqi Zhan^*, Changye Li^*, Zhijian Liu, Yao Lu, Yi Wu, Song Han, Ligeng Zhu
• (arXiv Preprint, 2025) Learning Personalized Driving Styles via Reinforcement Learning from Human Feedback
  Derun Li^*, Changye Li^*, Yue Wang^*, Jianwei Ren^*, Xin Wen, Pengxiang Li, Leimeng Xu, Kun Zhan, Peng Jia, Xianpeng Lang, Ningyi Xu, Hang Zhao
• (ICML 2025 Poster) SAE-V: Interpreting Multimodal Models for Enhanced Alignment
  Hantao Lou^*, Changye Li^*, Jiaming Ji, Yaodong Yang
• (ACL 2025 Best Paper Award) Language Models Resist Alignment: Evidence From Data Compression
  Jiaming Ji^*, Kaile Wang^*, Tianyi Qiu^*, Boyuan Chen^*, Jiayi Zhou^*, Changye Li, Hantao Lou, Josef Dai, Yunhuai Liu, Yaodong Yang
• (AAAI 2025 Poster) Towards efficient collaboration via graph modeling in reinforcement learning
  Wenzhe Fan, Zishun Yu, Chengdong Ma, Changye Li, Yaodong Yang, Xinhua Zhang

Projects

• Humanize: the agent flow system for token maxxing
  June 2026 – Present

  Core contributor, working on workflow infrastructure design, workflow design and optimization, and theoretical modeling.

• AReaL: Ant Reasoning Reinforcement Learning for LLMs
  June 2025 – October 2025

  Contributed to building training and inference infrastructure for vision-language reasoning models.

Honors & Awards

• Peking University Freshman Scholarship (2022)
• Dean’s Scholarship, Institute for Artificial Intelligence, Peking University (2025)

Service

• Reviewer: ICML 2025; ICLR 2026, 2027; NeurIPS 2024, 2025, 2026.

Language Proficiency

• Chinese: Native
• English: Advanced
• Hokkien: Intermediate
• Italian, Spanish: Basic