About Me

My name is Chenyue (Jim) Li. I’m a PhD student in the Department of Computer Science and Engineering at HKUST, under the supervision of Prof Binhang YUAN. My research interest primarily focuses on AI for Science(AI4S), aiming to boost the scientific discovery by leveraging the power of AI. Additionally, I also have complementary background in database, distributed Systems, computer vision, cybersecurity and full-stack development.

Education

• Hong Kong University of Science and Technology (HKUST) (2024 - ongoing)
  • PhD student in the Department of Computer Science and Engineering
• University of Toronto - St. George (2019 - 2024)
  • Honours Bachelor of Science with High Distinction
  • Computer Science Specialist
  • Dean’s List Scholar

Experience

• Voithos Labs (Sep 2023 - April 2024)
  • Software Engineer · Part-time
  • Developed alpha and beta versions for VNOTE, a super-contextual note-taking application with intelligent assistants. Reference: https://www.voithoslabs.com/.
• Huawei Technologies Canada Co., Ltd. (May 2022 - Sep 2023)
  • Assistant Engineer in Distributed Data Storage and Management Lab (R&D) · Co-op
  • Contributed to the development and optimization of MindSpore Pandas. Reference: https://www.mindspore.cn/en.
  • Contributed to the distributed computing engine in Jiutian (Huawei Analytics Engine) and GaussDB by researching and enhancing BSP and MPP execution modes in C++.
• Meonc Studio (Nov 2016 - Aug 2021)
  • Studio Founder · Freelance

Publications

• Zipeng Qiu, Chenyue Li, You Peng, Guangxin He, Binhang Yuan, and Chen Wang. “TQA-Bench: Evaluating LLMs for Multi-Table Question Answering.” IEEE Transactions on Big Data, accepted. (2026)

• Chenyue Li^*, Hyeonjae Kim^*, Wen Deng, Mengxi Jin, Wen Huang, Mengqian Lu, and Binhang Yuan. “ClimateAgent: Multi-Agent Orchestration for Complex Climate Data Science Workflows.” Transactions on Machine Learning Research (TMLR), April 2026.

• Chenyue Li^*, Wen Deng^*, Zhuotao Sun, Mengxi Jin, Hanzhe Cui, Han Li, Shentong Li, Man Kit Yu, Ming Long Lai, Yuhao Yang, Mengqian Lu, and Binhang Yuan. “S2SServiceBench: A Multimodal Benchmark for Last-Mile S2S Climate Services.” arXiv preprint arXiv:2602.14017. (2026)

• Chenyue Li, Wen Deng, Mengqian Lu, and Binhang Yuan. “AtmosSci-Bench: Evaluating the Recent Advance of Large Language Model for Atmospheric Science.” In The Thirty-ninth Annual Conference on Neural Information Processing Systems, Datasets and Benchmarks Track. (NeurIPS 2025)

• Lujia Zhang, Yurong Song, Hanzhe Cui, Mengqian Lu, Chenyue Li, Binhang Yuan, Bin Wang, Upmanu Lall, and Jing Yang. “Foundation Models as Assistive Tools in Hydrometeorology: Opportunities, Challenges, and Perspectives.” Water Resources Research 61, no. 4 (2025): e2024WR039553.

• David Anugraha, Genta Indra Winata, Chenyue Li, Patrick Amadeus Irawan, and En-Shiun Annie Lee. “ProxyLM: Predicting Language Model Performance on Multilingual Tasks via Proxy Models.” In Findings of the Association for Computational Linguistics: NAACL 2025, pp. 1981–2011. (2025)

• Guangxin He, Zonghong Dai, Jiangcheng Zhu, Binqiang Zhao, Qicheng Hu, Chenyue Li, You Peng, Chen Wang, and Binhang Yuan. “Zero-Indexing Internet Search Augmented Generation for Large Language Models.” arXiv preprint arXiv:2411.19478. (2024)

^* Equal contribution.

Research Projects & Activities

• Industrial AI Model Integration and Standardization (2026 - ongoing)
  • Research participant
  • Contribute to the technical roadmap and evaluation design for industrial semantic parsing, with research on domain-enhanced Text-to-SQL, agent skill orchestration, multi-source evidence fusion, and memory-augmented querying.
• S2S-HazardAgent: Localized Multi-Hazard Subseasonal Early Warning Support for Brazil (2026)
  • Project Lead · 2026 MAZU Global Intelligence Innovation Application Challenge
  • Led a team to design and develop an AI agent system for localized multi-hazard subseasonal risk analysis and early warning support in Brazil. Coordinated and integrated the application solution report, product prototype, model design, and presentation materials.
  • Selected for the 2026 World Artificial Intelligence Conference (WAIC) showcase. Prof. Mengqian Lu presented the system in a keynote at the meteorology session.
• SEPRESS Climate Intelligence Agent (2025 - ongoing)
  • Research Developer · SEPRESS Prediction and Services System, HKUST
  • Design and develop a climate intelligence agent, integrating climate products, forecast reports, and domain-specific APIs as agent tools to support natural language interaction, report retrieval, multi-step task execution, and interpretation of climate information.
  • Developed within SEPRESS, an HKUST-led programme endorsed by UNESCO as part of the International Decade of Sciences for Sustainable Development.
• Digital Twin-Enabled Wave Overtopping Prediction and Analysis under Storm Surges (2025)
  • Research course project team member · CIVL 5220, Hong Kong University of Science and Technology (HKUST)
  • Contributed to the development of a multi-agent prototype integrating meteorological data acquisition, wave overtopping simulation using smoothed particle hydrodynamics (SPH), 3D visualization, and evidence-based risk report generation.
• Exploring Image Alignment Methods for User Intent in Visual Generation
  • Course project team member · COMP5421, Hong Kong University of Science and Technology (HKUST)
  • Explored user-intent alignment through a ComfyUI and ControlNet workflow, comparing prompt-only and conditioned generation in qualitative case studies of pose, sketch, depth, and texture control.
  • Contributed to Illustrious-Agent, an OpenPose-based workflow combining iterative prompt optimization, evaluator feedback, context memory, and human review.
• Survey of Hybrid Search in Vector Databases
  • Co-author · COMP5311 course survey, Hong Kong University of Science and Technology (HKUST)
  • Co-authored a survey of vector similarity search with predicate filtering, categorizing methods into naive processing, fused indexing, and fused distance. Compared reported trade-offs in accuracy, efficiency, and scalability, and discussed future research directions.
• Multi-LLM Interaction
  • Joint course presentation · COMP6211J, Hong Kong University of Science and Technology (HKUST)
  • Co-presented a literature review of RLHF, DPO, Mixture-of-Agents, and Archon, discussing model collaboration, architecture search, and directions for query-specific coordination.

Awards & Honors

• RedBird Academic Excellence Award, Hong Kong University of Science and Technology (2025–2026)
• Dean’s List Scholar, University of Toronto (2023)
• Dean’s List Scholar, University of Toronto (2021–2022)
• 4th Place, THE Hack Hackathon, Shanghai (2019)
  • Project: AI Tutor: Integrated Learning Platform

Acknowledgements

• David Anugraha, Vishakh Padmakumar, and Diyi Yang. “SparkMe: Adaptive Semi-Structured Interviewing for Qualitative Insight Discovery.” arXiv preprint arXiv:2602.21136. (2026) Acknowledged contributor: Chenyue Li.

Background & Interests

• AI for Science (AI4S)
• Large Language Model (LLM) & Machine Learning (ML) & Natural Language Processing (NLP)
• Database & Distributed Systems
• Cybersecurity
• Full Stack Development
• Computer Graphics