About Me

My name is Zheyuan (Jason) Zhang, a third-year Ph.D. student in the Department of Computer Science and Engineering at the University of Notre Dame, advised by Prof. Yanfang (Fanny) Ye.

My research focuses on post-training strategies for large language models, with an emphasis on agent optimization and self-evolving agent behaviors. Specifically, my work spans three areas:

• Dynamic Multi-agent System Design: prompt optimization (EACL’26), self-evolving agents (Arxiv; EACL’26), agent collaboration (NeurIPS’25; Arxiv)
• Retrieval-augmented Reasoning: (KDD’24; KDD’25; ACL’25; EACL’26), for data analysis with long context (Arxiv).
• Uncertainty-aware Reinforcement Learning: intention clarifying, uncertainty beyond entropy, uncertainty in VLA.

For more details, please check my CV here.

📚 Selected Publications

You can find the full list of my articles on my Google Scholar profile.

AgentRouter: A Knowledge-Graph-Guided LLM Router for Collaborative Multi-Agent Question Answering [Arxiv Oct.]

Zheyuan Zhang*, Kaiwen Shi*, Zhengqing Yuan, Zehong Wang, Tianyi Ma, Keerthiram Murugesan, Vincent Galassi, Chuxu Zhang, Yanfang Ye

MAPRO: Recasting Multi-Agent Prompt Optimization as Maximum a Posteriori Inference [EACL'26]

Zheyuan Zhang, Lin Ge, Hongjiang Li, Weicheng Zhu, Chuxu Zhang, Yanfang Ye

NG-Router: Graph-Supervised Multi-Agent Collaboration for Nutrition Question Answering [EACL'26]

Kaiwen Shi*, Zheyuan Zhang*, Zhengqing Yuan, Keerthiram Murugesan, Vincent Galass, Chuxu Zhang, Yanfang Ye

NGQA: A Nutritional Graph Question Answering Benchmark for Personalized Health-aware Nutritional Reasoning [ACL'25]

Zheyuan Zhang*, Yiyang Li*, Nhi Ha Lan Le*, Zehong Wang, Tianyi Ma, Vincent Galassi, Keerthiram Murugesan, Nuno Moniz, Werner Geyer, Nitesh V Chawla, Chuxu Zhang, Yanfang Ye

MOPI-HFRS: A Multi-objective Personalized Health-aware Food Recommendation System with LLM-enhanced Interpretation [KDD'25]

Zheyuan Zhang, Zehong Wang, Tianyi Ma, Varun Sameer Taneja, Sofia Nelson, Nhi Ha Lan Le, Keerthiram Murugesan, Mingxuan Ju, Nitesh V Chawla, Chuxu Zhang, Yanfang Ye

Diet-ODIN: A Novel Framework for Opioid Misuse Detection with Interpretable Dietary Patterns [KDD'24]

Zheyuan Zhang*, Zehong Wang*, Shifu Hou, Evan Hall, Landon Bachman, Vincent Galassi, Jasmine White, Nitesh V. Chawla, Chuxu Zhang, Yanfang Ye

💻 Work Experience

Amazon Science, Seattle, WA, USA

• Aug. 2025 – Dec. 2025, Applied Scientist (Part-time)
• May. 2025 – Aug. 2025, Applied Scientist (Intern)

IBM Research, Yorktown Heights, NY, USA

• Jun. 2024 – May. 2025, Student Researcher (Part-time)

Coleridge Initiative, New York, NY, USA

• Oct. 2021 – Mar. 2023, Research Scientist
• Aug. 2020 – Oct. 2021, Associated Research Scientist
• Oct. 2019 – Aug. 2020, Junior Research Scientist (Intern)

📖 Education

• 2023.07 - present, Ph.D., University of Notre Dame
• 2019.08 - 2020.08, M.S., New York University
• 2015.09 - 2019.06, B.S., Wuhan University

📝 Mentorship

I’ve been fortunate to mentor and work alongside many brilliant minds:

Current Mentees

• Han Bao, Senior Undergraduate, Sichuan University (Sept. 2025 – present)
• Xiaoyu Ma, M.S. Student, National University of Singapore (Oct. 2025 – present)

Former Mentees

• Nhi Le, M.S. Student, Brandeis University (Jun. 2024 – Jan. 2025)
• Vinny Galassi, Junior Undergraduate, University of Notre Dame (2023 – 2025)
• Sofia Nelson, Senior Undergraduate, University of Notre Dame (2024)
• Varun Taneja, Senior Undergraduate, University of Notre Dame (2024)
• Evan Hall, Senior Undergraduate, University of Notre Dame (2023 – 2024)
• Landon Bachman, Senior Undergraduate, University of Notre Dame (2023 – 2024)
• Jasmine White, Senior Undergraduate, Purdue University (2023 – 2024)

💬 Invited Talks

• MOPI-HFRS (Sept. 2024) @ IBM Research, Thomas J. Watson Research Center, Yorktown Heights.
• LLM Agents for Code Generation (Jun. 2025) @ Amazon, Seattle.

📢 Contact

• Email: zzhang42 [at] nd [dot] edu
• Office: 247B Fitzpatrick Hall of Engineering
• Location: University of Notre Dame, Notre Dame, IN 46565

Feel free to reach out via email if you’re interested in our research or potential collaborations.