I am a PhD student at the Hong Kong University of Science and Technology (Guangzhou), starting from Fall 2024. I am honored to be supervised by Professor Yuyu Luo. Previously, I obtained my bachelor’s degree from the School of Information, Renmin University of China. My research interest primarily focuses on LLM Based Agents, with a particular emphasis on reasoning ablity and the paradigm of constructing agents that are adaptable to multiple environments.
I am building a community called MindForge, aimed at bringing together researchers and engineers interested in intelligence to explore autonomous and efficient agents across different environments. If you are interested, feel free to reach out to me.
You can find my CV here: didi’s Curriculum Vitae.
🔥 News
- 2025.02 We proposed AlphaSQL, the o1 moment for NL2SQL!
- 2025.02 We proposed SPO (Self-Supervised Prompt Optimization) and AoT (Atom of Thoughts for Markov Test-Time Scaling)
- 2025.02 AFLOW was accepted by ICLR 2025 as an Oral Presentation (Top 1.8%)
- 2024.08 I joined the DataLab at HKUST(GZ) as a PhD student, advised by Professor Yuyu Luo.
- 2024.07 We proposed MobileExperts, a Multi-Agent system in GUI environments.
- 2024.06 MathAI won third place in the 2024 Alibaba Global Mathematics Competition AI Challenge, thanks to SuperCarryMan!
- 2024.02 Participated in my first academic paper, delighted to collaborate with the MetaGPT community to publish Data Interpreter.
- 2023.12 Participated in the Baidu Hackathon with my homies and won second place.
📝 Selected Publications
AFlow: Automating Agentic Workflow Generation
Jiayi Zhang, Jinyu Xiang, Zhaoyang Yu, Fengwei Teng, Xionghui Chen, Jiaqi Chen, Mingchen Zhuge, Xin Cheng, Sirui Hong, Jinlin Wang, Bingnan Zheng, Bang Liu, Yuyu Luo, Chenglin Wu
We introduces AFlow, an automated framework that reformulates workflow optimization as a search problem over code-represented workflows, using Monte Carlo Tree Search to efficiently explore and refine workflows through code modification and execution feedback. By leveraging this approach, AFlow achieves superior performance compared to state-of-the-art baselines across multiple benchmarks, while also enabling smaller models to outperform larger ones at a fraction of the cost.
📖 Educations
- 2024.08 - Present, The Hong Kong University of Science and Technology (Guangzhou), Information Hub.
- 2020.09 - 2024.06, Renmin University of China, Information School.
💻 Internships
- 2024.7 - Present, Research Intern, DeepWisdom, China.
- 2024.6 - 2025.2, Research Intern, Thin Red Line, China.
- 2023.12 - 2024.5, Research Intern, Lenovo Research AI Lab, China.
- 2023.08 - 2023.12, Algorithm intern, ZhiPuAI ,China.
- 2023.07 - 2023.08, Algorithm intern,CYOU, China.
🎯 Competitions and Side Projects
- 2024.6, 2024 Alibaba Global Mathematics Competition AI Challenge, Top 1% (3/563), Team Leader
- 2024.4, MetaGPT Android Environment, Core Contributor
- 2023.12, Baidu & FounderPark AGI Hackathon, Second Place Award, Team Leader
- 2023.9, MetaGPT Mini Hackathon, Best Contributor
- 2023.8, AWS GenAl Talent Program, 2 out of 4 awards, Team member
🎖 Honors and Awards
- 2022.10 The First Class Academic Progress Scholarship, Renmin University of China.
- 2022.10 The First Class Social Work and Leadership Scholarship, Renmin University of China.
- 2023.10 The First Class Outstanding Student Leadership Scholarship, Renmin University of China.
👥 Leaderships
- 2023.07 - 2024.06, RUC 36th Student Union Committee, Member.
- 2023.04 - 2024.03, RUC 45th School of Information Student Union, Chairman.
- 2022.07 - 2023.07, RUC lth Mingli Academy Student Union, Chairman.
🔈 invited-talks
- 2024.10, GraphTalk: Automating Effective Agentic Workflow Generation, Ant Group
- 2025.3, ThinkZo Workshop, From OpenManus to Foundation Agents, HUAWEI