Biography 🙋

中文/EN

I am a second-year Ph.D. student at the Department of Data Science & Artificial Intelligence (DSAI) / Department of Applied Mathematics (AMA), the Hong Kong Polytechnic University (PolyU) PolyU Logo. I am fortunately supervised by Prof. Huang Jian, Prof. Han Ruijian, and Prof. Yuan Yancheng. I will be a visiting researcher at University of North Carolina at Chapel Hill (UNC) UNC Logo from 2027.

Before that, I worked as a research assistant at Research Center for the Mathematical Foundations of Generative AI (CMFAI), supervised by Prof. Huang Jian. I obtained my Master's degree with Distinction Honor in Data Science in 2024, under supervision of Prof. Jiang Binyan. I got my Bachelor's degree in Computer Science in 2022, and obtained the National Scholarship.

I am passionate about artificial intelligence, especially large language models (LLMs), agents, and reinforcement learning related research and applications. Feel free to reach out for discussion and collaboration.

Research Interests 💡

  • Agentic AI: Data Science Agent, Multi-agent System, Retrieval Augmented Generation, Benchmarks.
  • Large Language Model: Supervised Fine-tuning and Reinforcement Learning.
  • AI4Science: Agent4Science, Health Informatics, Medical Image, Medical Language Model.

News & Moments 📢

April 20, 2026
LAMBDA Web App is now online and free to use: https://lambda.com.ai . Welcome to try it out and give us feedback!
User manual: https://www.lambda.org.ai/manual/
July 20, 2026
We launch DSArena, an arena for evaluating large language models on data science tasks: https://lambda.com.ai/dsarena . You can select DSArena in the LAMBDA Web App to perform data analysis tasks. With a single instruction, multiple models will work for you simultaneously.
For more details, please visit: the DSArena blog . Come and vote for the model best at data analysis!
May 4, 2026
I passed my Confirmation of Registration in PolyU and became a Ph.D. candidate.
April 15, 2026
All discussion papers and rejoinder of LAMBDA have been published. We thank Prof. Hongtu Zhu for his organization. We also thank Prof. David Donoho, Prof. James Zou, Prof. Xihong Lin, Prof. Xiao-Li Meng, Prof. Fan Zhou, Dr. Rui (Sammi) Tang, and other participants for their valuable comments. We have carefully considered these comments and provided detailed responses (rejoinder paper).
March 05, 2026
We released paper DARE: Aligning LLM Agents with the R Statistical Ecosystem via Distribution-Aware Retrieval, an embedding model for improving LLM Agents on R package retrieval and downstream statistical analysis tasks. DARE outperforms open-sourced embedding models on R retrieval with higher efficiency and accuracy.
Jan 20, 2026
We released paper DSAEval: Evaluating Data Science Agents on a Wide Range of Real-World Data Science Problems, introducing a large-scale benchmark with 641 real-world data science tasks across multimodal data and interactive analytical workflows, enabling systematic evaluation of modern data science agents.
Aug 22, 2025
Our paper A Survey on Large Language Model-based Agents for Statistics and Data Science has been accepted by TAS (The American Statistician). And It is selected with discussion 🎉.
May 16, 2025
Paper LAMBDA: A Large Model Based Data Agent accepted by top journal JASA 🎉. And it is our great honor that the paper is selected with discussion (Only 2) and will be presented at JSM 2025. We are especially privileged to learn that Prof. David Donoho will serve as one of the discussants for our work.
Aug 30, 2024
Registered as a Ph.D. student at Hong Kong Polytechnic University.
Jul 15, 2024
Graduated with Distinction Honor 🥇 from Msc in Data Science & Analytics, PolyU.
Dec 2023
My fans exceed 1000 on CSDN 🔥.

Publications 📰

(View all on Google Scholar )

† Equal contribution. * Corresponding author.

Asymmetric retrieval geometry
When Does Dense Retrieval Need Asymmetric Geometry? Local Manifold Risk and Uncertainty-aware Selection
Maojun Sun, Ruijian Han*, Binyan Jiang, Defeng Sun, Yancheng Yuan*, Jian Huang*.
2026
AutoXRD
AutoXRD: Autonomous Large Language Model Agents and Evaluation for Powder X-Ray Diffraction Analysis
Yuetong Wu†, and Maojun Sun†*
2026
Statistical–visual verification framework
Do the Numbers Match the Plot? Statistical–Visual Verification for Automated Exploratory Data Analysis
Zixun Fu†, Jian Huang†,∗, Maojun Sun†, and Yancheng Yuan†,∗
Under Review, 2026

Experiences 🚀

Hong Kong Polytechnic University Feb 2024 - Aug 2024
Research Assistant & Project Assistant
Research and develop on LLM for data analysis. Develop systems for IOR, CMFAI, RCNA, RCQF.
AI Lab, Hong Kong Hospital Authority Jul 2023 - Dec 2023
Student Researcher
Research and develop on LLM in diagnostic systems (fine-tuning, evaluation, prompt engineering).
Bacara Energy Technology Jun 2022 - Aug 2022
Image Algorithm Intern
Research and develop on intelligent inspection solutions for wind power drones.
DXC Technology Nov 2021 - Jun 2022
AI Engineer
Develop on web robot. Research on algorithm of resume content classification and extraction.

Awards 🏅

National Scholarship of China 12/2020
Outstanding Graduation of Zhejiang Province 06/2022
Government Scholarship of Zhejiang Province 12/2021
PolyU Research Postgraduate Scholarship 09/2024
Elite Scholarship × 2 2020, 2021
First Class Scholarship × 5 2018-2021
Winning Prize, DJI RoboMaster Intelligent Perception 12/2022
Second Prize of National Artificial Intelligence & Innovation Competition 05/2021
Merit Student Award × 7 2018 & 2019 & 2020 & 2021

Teaching 👨‍🏫

Teaching Assistant, DSAI1102 Data Analytics Fundamentals, PolyU 25/26 S2
Teaching Assistant, DSAI5101 Statistical Data Mining, PolyU 25/26 S1
Teaching Assistant, Mathematics Learning Support Centre, PolyU 24/25 S1

Talks & Conferences 🎙️

Agentic LLMs for Statistics: The LAMBDA System and Beyond @ Joint Statistical Meetings 2026 (JSM2026), Boston, Massachusetts, by Prof. Jian Huang. Aug 2026
Agentic LLMs for Statistics: The LAMBDA System and Beyond @ Korean Statistical Society Summer Conference 2026, Gyeongju, South Korea, by Prof. Yancheng Yuan. Jul 2026
Towards Agentic AI for Automated Data Science and Statistics @ PolyU PhD Confirmation of Registration. May 2026
LAMBDA: A Large Model Based Data Agent @ Joint Statistical Meetings 2025 (JSM2025), by Prof. Jian Huang. Aug 2025
LAMBDA: A Large Model Based Data Agent @ Seminar of Mathematical Foundations of AI, Tianyuan Mathematics Research Center, Kunming, Yunnan Sep 2024
Large Language Models: Principles, Evolution, and Applications @ PolyU Summer School Jun 2024

Professional Skills 🪀

AI & ML: LLM Fine-tuning, Image Classification, Data Mining, Target Detection, Image Segmentation, etc.

Programming: Python, Java, SQL, HTML/JS/CSS, C, etc.

Development: FastAPI, Flask, SpringBoot, SpringCloud, Vue, Nginx, Git, Docker, AWS, Aliyun, etc.

Big Data: MySQL, Redis, Hadoop, Spark, etc.

Others 💌

🎓 Research interests, please contact me by mj.sun@connect.polyu.hk