DATA · AGENTS · WORLD MODELS

Scalable data synthesis
for agent training.

About me

I am Meihao Fan (范梅浩), a Ph.D. student in Computer Application Technology at Renmin University of China, advised by Prof. Ju Fan and Prof. Xiaoyong Du. I completed my B.S. at Chongqing Jiaotong University in 2023.

My research focuses on scalable data synthesis for agent training. I develop methods for task generation, interaction-trajectory synthesis, and execution-based data verification, and use the resulting data for curriculum learning and multi-turn reinforcement learning. My work connects data construction to measurable improvements in agent capability and generalization. I am extending this research to embodied data curation, action-conditioned world models, and evaluation-driven data evolution.

I am seeking a Summer 2027 U.S. research internship in embodied AI data, world models, and multimodal agents.

CV · GitHub · Google Scholar · LinkedIn

Research Experience

  • EvoPhys, Beijing, China — Research Intern, World Models — Sept 2026 – Present
    • Lead research on world model training, spanning interactive-memory evaluation, scalable multimodal data annotation, and evaluation-driven recursive self-improvement (RSI).
  • ByteDance, Beijing, China — Research Intern, LLM Agent Systems & Reinforcement Learning — March 2025 – June 2026
    • Led DeepPrep, covering synthetic training tasks, an executable data-preparation environment, and progressive SFT/RL; led research on agent harness self-evolution from execution trajectories.
  • Renmin University of China, Beijing, China — Research Assistant — Sept 2023 – March 2025
    • Led AutoPrep and BATCHER research on task-aware data preparation, multi-agent orchestration, demonstration selection, and cost-efficient LLM inference.

Selected Research

More projects

Selected Publications

  • Meihao Fan, Ju Fan, Yuxin Zhang, Shaolei Zhang, Xiaoyong Du, Jie Song, Peng Li, Fuxin Jiang, Tieying Zhang, Jianjun Chen. DeepPrep: An LLM-Powered Agentic System for Autonomous Data Preparation VLDB 2026. Paper

  • Meihao Fan, Ju Fan, Nan Tang, Lei Cao, Guoliang Li, Xiaoyong Du. AutoPrep: Natural Language Question-Aware Data Preparation with a Multi-Agent Framework VLDB 2025. Paper · Slides

  • Meihao Fan, Xiaoyue Han, Ju Fan, Chengliang Chai, Nan Tang, Guoliang Li, Xiaoyong Du. Cost-Effective In-Context Learning for Entity Resolution: A Design Space Exploration ICDE 2024. Paper · Slides

  • Shaolei Zhang, Ju Fan, Meihao Fan, Guoliang Li, Xiaoyong Du. DeepAnalyze: Agentic Large Language Models for Autonomous Data Science ICML 2026. Paper · Code

  • Yuxin Zhang, Meihao Fan, Ju Fan, Mingyang Yi, Yuyu Luo, Jian Tan, Guoliang Li. Reward-SQL: Boosting Text-to-SQL via Stepwise Execution-Aware Reasoning and Process-Supervised Rewards SIGMOD 2026. Paper

  • Yuxin Zhang, Ju Fan, Meihao Fan, Shaolei Zhang, Xiaoyong Du. CoDA-Bench: Can Code Agents Handle Data-Intensive Tasks? ICML 2026. Paper

  • Chao Deng, Ju Fan, Yuyu Luo, Qinliang Xue, Meihao Fan, Yuxin Zhang, Min Zhang, Xiaofeng Jia, Jing Zhang, Xiaoyong Du. TACO: A Benchmark for Open-Domain Text-to-SQL with Ambiguous and Cross-Database Queries VLDB 2026. Paper

Manuscripts

  • Meihao Fan, Shaolei Zhang, Ju Fan, Peng Li, Jie Song, Jianjun Chen. EvoCost: Harness Self-Evolution for Cost-Efficient LLM Agents Under Review at ICLR 2027.

  • Gaoyuan Li, Meihao Fan, Yizhe Liu, Shaolei Zhang, Ju Fan, Siyi Wang, Jiaheng Hou, Xudong Weng, Honghan Tian, Zang Li. SkillAdam: Stable and Efficient Skill Evolution for Agents Under Review at ICLR 2027. Preprint · Code

  • Yuxin Zhang, Ju Fan, Meihao Fan, Shaolei Zhang, … Multi-Agent World: Scaling Multi-Agent Orchestration Training via Recursive Environment Improvement Under Review at ICLR 2027.

Selected Awards

  • National Scholarship (Ph.D.), 2025 — the college’s only second-year Ph.D. recipient.
  • National Scholarship (B.S.), 2023 — the college’s only recipient.
  • Mingde Scholarship Nomination Award, Chongqing Jiaotong University, 2023 — university top 20.