Curriculum Vitae

Associate Researcher at Beihang University, Hangzhou International Innovation Institute. Contact: huiwenw0820@outlook.com.

Experience

2026 –
Associate Researcher Beihang University, Hangzhou International Innovation Institute · Hangzhou, Zhejiang
- Trustworthy large language models and AI for science.
- Developing mathematically grounded privacy-preserving algorithms for large-scale foundation models, with domain-driven applications in geophysics (seismic imaging) and astronomy (galaxy simulation for CSST).
2022–26
Senior Researcher Zhejiang Laboratory · Hangzhou, Zhejiang
- Trustworthy large language models and AI for scientific discovery.
- Bridged numerical analysis, differential privacy, and scientific machine learning to enable collaborative AI.
2020–22
Senior Algorithm Engineer Ant Group · Hangzhou, Zhejiang
- Trustworthy federated learning.
- Developed fast and robust optimization methods for privacy-preserving federated learning and transfer learning.

2019
Ph.D. in Mathematics University of California, Irvine · Irvine, California
- Thesis: Randomized Fast Solvers for Linear and Nonlinear Problems in Data Science.
- Advisor: Prof. Long Chen. GPA 3.97/4.0.
2013
B.A. in Mathematics Sichuan University · Chengdu, Sichuan
- Graduated with Honors. GPA 3.75/4.0.
- Top-notched Student Scholarship, 2009–2013.

2026–
Matrix Theory and Applications I M571011S03 · Graduate · 16 weeks · 32 contact hours

2026–31
AI-Powered Galaxy Simulation and Evolution of Milky Way Analogs for the Chinese Space Station Telescope (CSST) National Key R&D Program, Engineering Sciences and Interdisciplinary Integration · 20,000,000 CNY
Core Member
2023–26
Key Technologies and Applications of Large-Scale Distributed Trustworthy Intelligent Computing Based on Blockchain National Key R&D Program, Blockchain Key Special Project · 19,982,000 CNY
Participant
2023–24
Zhejiang Lab Talent Fund for Young Professionals 200,000 CNY
Principal Investigator
2024
Young Talent Nurturing Project, Zhejiang Lab Selected, August – November 2024

2026
Synergizing Multigrid Algorithms with Vision Transformer: A Novel Approach to Enhance the Seismic Foundation Model AAAI, oral
2024
Advancing Large Language Model Privacy and Efficiency in Federated Learning: Empirical Improvements Chinese Academy of Sciences, invited
2024
A Variational Approach to Personalized Federated Learning and its Improvement PKAW, oral
2022
Privacy-preserving techniques in Transfer and Federated Learning Chinese Academy of Sciences, invited
2022
Differential Private Knowledge Transfer for Privacy-Preserving Cross-Domain Recommendation The Web Conference (WWW)
2017
A Preconditioner based on Non-uniform Row Sampling for Linear Least Squares Problems Southern California Applied Mathematics Symposium (SOCAMS)

Privacy & FL: Differential privacy (RDP/GDP accounting), federated optimization, secure aggregation
Optimization: Stochastic gradient methods, preconditioning, multigrid algorithms, variational inference
Domains: Seismic imaging, galaxy simulation, cross-domain recommendation, large language models
Tools: PyTorch, TensorFlow, Olmo, CUDA

PC Member: AAAI'23–'26, AAAI-AIA'26, PRICAI'23–'25, INCRYPT'23
Reviewer: AISTATS'22–'24, WWW'25, ACM MM'24–'25, ICME'24–'26, ICASSP'25, IEEE TIFS
Session Chair: IJCAI-ECAI'22, PKAW'24