🌐 中文版

Jinhao Dong, Ph.D., is a lecturer in the Key Laboratory of Data Engineering and Knowledge Engineering (MOE) and the School of Information, Renmin University of China (RUC), and is selected into the Wu Yuzhang Young Talents Program (吴玉章青年英才) of RUC. He joins the research group led by Professor Xiaoyong Du and Professor Wei Lu. He is also the technical advisor of Xiaomi’s LLM Core Team, and a core contributor to the MiMo series of foundation models — he is primarily responsible for code-related LLM development, including pretraining, reinforcement learning, Code Agent, browser agent, and the OpenClaw general agent platform. The MiMo-V2-Pro model that he core-contributed ranks #8 globally and #2 in China on the Artificial Analysis Intelligence Index, and has been the #1 model by call volume on OpenRouter. He took the lead in establishing the RUC–Xiaomi Joint Key Laboratory for Foundation Large Language Models (the first joint key laboratory under RUC’s new policy) and serves on its Technical Committee.

He has published 17 papers at CCF-A top-tier conferences and journals (11 as first / corresponding author), covering ICML, NeurIPS, KDD, ICSE, ASE, ISSTA, FSE, etc.

He obtained his Ph.D. from the School of Computer Science, Peking University in 2025, under the supervision of Professor Dan Hao, and his Bachelor’s degree from the School of Computer Science and Technology, Xidian University. He was a visiting scholar in National University of Singapore (NUS) for one year, under the supervision of Prof. Jun Sun, Prof. Yun Lin, and Prof. Jin Song Dong.

His current research focuses on Large Language Models (LLMs), building the next-generation trustworthy, general, and autonomous LLM & agent systems. Specifically:

Code LLM — code pretraining/posttraining and Code Agent reinforcement learning
LLM Agent — planning & memory, tool use, browser agent, multi-agent collaboration, agentic interaction data synthesis
LLM Infra — long-context inference acceleration (KVCache compression & quantization), robust fine-tuning
General Agent Platform — OpenClaw general agent system, large-scale real-world user interaction data & training

If you're interested in our research group or in pursuing an internship at Xiaomi's LLM Core Team, feel free to email me at dongjinhao@ruc.edu.cn.

📖 Educations

2020.09 - 2025.07, Peking University, Ph.D. in Computer Science, supervised by Prof. Dan Hao
2023.11 - 2024.11, National University of Singapore (NUS), Visiting Scholar in Computer Science, supervised by Prof. Jun Sun, Prof. Yun Lin, and Prof. Jin Song Dong
2016.09 - 2020.07, Xidian University, Bachelor in Software Engineering

💼 Experience

2025.08 - Present, Renmin University of China, School of Information & Key Lab of Data Engineering and Knowledge Engineering (MOE), Lecturer
2025.08 - Present, RUC–Xiaomi Joint Key Laboratory for Foundation LLMs, Technical Committee Member
2024.11 - Present, Xiaomi LLM Core Team, Technical Advisor

📝 Publications

Total: 17 papers at CCF-A top venues, 11 as first / corresponding author. (* denotes corresponding author)

[Technical Report] MiMo: Unlocking the Reasoning Potential of Language Model – From Pretraining to Posttraining [PDF] [GitHub]

Core contributor, responsible for code pre-training and reinforcement learning. Processed the code pre-training and post-training data, and proposed a test-difficulty–driven code reward to mitigate the sparse-reward problem in code reinforcement learning.
[Technical Report] MiMo-VL Technical Report [PDF] [GitHub]

Contributor
[Technical Report] MiMo-Audio: Audio Language Models are Few-Shot Learners [PDF] [GitHub]

Contributor
[ICML’26] RaBitQCache: Rotated Binary Quantization for KVCache in Long Context LLM Inference

Wenhao Li, Jinhao Dong*, Hailin Zhang, Wenhang Shi, Wei Lu, Xiaoyong Du
[ICML’26] Training Prompt Matters: State-Adaptive Prompt Optimization for Robust Fine-Tuning (SAPO)

Wenhang Shi, Yiren Chen, Shuqing Bian, Zhe Zhao, Jinhao Dong*, Pengfei Hu, Wei Lu, Xiaoyong Du
[KDD’26] Scaling Agentic Capabilities via Grounded Interaction Synthesis (GAIS)

Wenhang Shi, Jinhao Dong*, Yiren Chen, Zhe Zhao, Shuqing Bian, Wei Lu, Xiaoyong Du
[KDD’26] BiVCoder: A Multi-Agent Framework for Code Generation via Bidirectional Code-Test Diagnosis

Xiaoyang Li, Jinhao Dong*, Wenhang Shi, Wei Lu, Xiaoyong Du
[EMNLP’26 Findings] GraP-Mem: Granularity-Aware Planning for Adaptive Memory Access in Long-Horizon LLM Agents

Qinyang Zeng, Jinhao Dong*, Wei Lu, Xiaoyong Du
[EMNLP’26 Findings] PersonaForge: Realistic Multi-Turn User Simulation for Agentic Systems

Hanglong Lv, Dawei Zhu, Lei Li, Bowen Ye, Huaqiu Liu, Yifan Song, Bofei Gao, Weimin Xiong, Jinhao Dong, Chenhong He, Lingpeng Kong, Qi Liu, Tong Yang, Fuli Luo
[ISSTA’25] ConTested: Consistency-Aided Tested Code Generation with LLM [PDF] [GitHub]

Jinhao Dong, Jun Sun, Wenjie Zhang, Jin Song Dong, and Dan Hao
[ASE’24] Revisiting the Conflict-Resolving Problem from a Semantic Perspective [PDF] [GitHub]

Jinhao Dong, Jun Sun, Yun Lin, Yedi Zhang, Murong Ma, Jin Song Dong, and Dan Hao
[ASE’23] Merge Conflict Resolution: Classification or Generation? [PDF] [GitHub]

Jinhao Dong, Qihao Zhu, Zeyu Sun, Yiling Lou, and Dan Hao
[ICSE’23] Revisiting Learning-based Commit Message Generation [PDF] [GitHub]

Jinhao Dong, Yiling Lou, Dan Hao, and Lin Tan
[ICSE’22] FIRA: Fine-Grained Graph-Based Code Change Representation for Automated Commit Message Generation [PDF] [GitHub]

Jinhao Dong, Yiling Lou, Qihao Zhu, Zeyu Sun, Zhilin Li, Wenjie Zhang, and Dan Hao
[ICSE’22] Recommending Good First Issues in GitHub OSS Projects [PDF] [GitHub]

Wenxin Xiao, Hao He, Weiwei Xu, Xin Tan, Jinhao Dong, and Minghui Zhou
[SCIS] What can we learn from quality assurance badges in open-source software? [PDF] [GitHub]

Feng Li, Yiling Lou, Xin Tan, Zhenpeng Chen, Jinhao Dong, Yang Li, Xuanzhi Wang, Dan Hao, and Lu Zhang
[ESEC/FSE’21] Boosting Coverage-Based Fault Localization via Graph-Based Representation Learning [PDF] [GitHub]

Yiling Lou, Qihao Zhu, Jinhao Dong, Xia Li, Zeyu Sun, Dan Hao, Lu Zhang, and Lingming Zhang
[ASE’20 NIER track] SRRTA: Regression Testing Acceleration via State Reuse [PDF] [GitHub]

Jinhao Dong, Yiling Lou, and Dan Hao
[ISSRE’20] Cost-effective testing of a deep learning model through input reduction [PDF] [GitHub]

Jianyi Zhou, Feng Li, Jinhao Dong, Hongyu Zhang and Dan Hao
[NeurIPS’19] MarginGAN: Adversarial Training in Semi-Supervised Learning [PDF] [GitHub]

Jinhao Dong, Tong Lin

🎖 Honors and Awards

2025: Selected into the Wu Yuzhang Young Talents Program (吴玉章青年英才), Renmin University of China
2023.12: Exceptional Award for Academic Innovation, Peking University
2022.12: Merit Student, Peking University
2022.12: Ubiquant Scholarship, Peking University
2021.12: Exceptional Award for Academic Innovation, Peking University
2021.12: Sanda Fighting Match, Bronze Medal, Peking University
2021.09: President Scholarship, Peking University
2020.09: President Scholarship, Peking University
2020.06: Outstanding Graduate, Xidian University
2019.11: President Fellowship, Xidian University
2018.10: International Collegiate Programming Contest (ICPC), Bronze Medal
2018.04: Mathematical Contest in Modeling (MCM), Meritorious Winner

💻 Services

Program Committee Member

2026: ICSE Research Track
2025: ASE Research Track
2024: ASE Research Track
2024: NeurIPS Main Conference
2023: PRDC Fast Abstract Committee

Reviewer

2024: Formal Aspects of Computing (FAC)
2024: Science China Information Sciences (SCIS)

🥊 Hobbies

Sanda (Chinese Kickboxing)
Breaking (Breakdance)
Hiking and Camping

🐱 My Cat

Meet 11 (Eleven), my American Shorthair kitty. He got his name because I brought him home in November, and because I’m a big fan of Stranger Things — Eleven is the show’s female lead.