Duc-Anh Nguyen

Thank you for visiting! My name is Duc-Anh Nguyen. You can call me Duc Anh for short. I’m an AI Research Resident at Qualcomm AI Research (formerly VinAI Research), advised by Dr. Toan Tran, Dr. Tung Pham, and Prof. Tan Nguyen.

Research Interest

My research focuses on uncovering the mathematical principles underlying foundation models, leveraging these insights to design efficient, interpretable, and robust LLMs and VLMs, with an emphasis on optimization, sparse modular architectures, and transfer learning.

News

  • [May 2026] πŸŽ‰ My two first-author papers at Qualcomm have been accepted at ICML 2026 workshops: "Selective Sinkhorn Routing for Improved Sparse Mixture of Experts" at AdaptFM and "Mixture-of-Control: State-Aware Fine-Tuning for Transformer-based Models" at CoLoRAI. See you in Seoul, South Korea this July πŸ‡°πŸ‡·!
  • [Mar 2026] πŸŽ“ I am thrilled to be admitted to Tufts University for a PhD in Electrical and Computer Engineering and National University of Singapore (NUS) for a PhD in Computer Science!
  • [Jan 2026] πŸŽ“ I am thrilled to be admitted to the University of Virginia as a PhD student in Computer Science, starting Fall 2026!
  • [Jan 2026] πŸ“‹ Our paper "Mixture-of-Control: State-Aware Fine-Tuning for Transformer-based Models" is approved for patent filing at Qualcomm. Paper coming soon!
  • [Jan 2026] πŸ“„ Our paper on "AI in Metal Additive Manufacturing" is published in Journal of Intelligent Manufacturing (Q1, IF: 7.4).
  • [Nov 2025] πŸ“‹ Our paper "Selective Sinkhorn Routing for Improved Sparse Mixture of Experts" is available on arXiV and is approved for patent filing at Qualcomm.
  • [Feb 2025] πŸ“ Our new preprint is on arXiv: "A Steepest Gradient Method with Nonmonotone Adaptive Step-sizes."
  • [Sep 2024] πŸ† I graduated as Valedictorian (Ranked 1st) from HUST with B.Sc in Mathematics and Informatics!
  • [Mar 2024] πŸ’Ό I joined VinAI Research (now Qualcomm AI Research) as an AI Research Resident.
  • [Nov 2023] πŸ₯‡ I won a Gold Medal at the International Youth Math Challenge!
  • [Nov 2023] πŸ… I was awarded the VIASM Scholarship from the Vietnam Institute for Advanced Study in Mathematics.
  • [Aug 2023] πŸ₯‰ I was awarded the Third Prize at VIASM Summer School "Research Experiences for Undergraduates".
  • [Jun 2022] πŸ₯‡ Our team won the Gold Medal at the Iranian Combinatorics Olympiad!
  • [Mar 2022] πŸ“š I started my research journey at HATLAB, HUST under the supervision of Dr. Thang Tran Ngoc.

Educations

Hanoi University of Science and Technology (HUST)

B.Sc in Mathematics and Informatics

πŸ† Valedictorian (Ranked 1st) CGPA: 3.92/4.00
Oct 2020 – Sep 2024
Hanoi, Vietnam

Thesis: Steepest Descent for Multi-Objective Optimization and Machine Learning Applications
Supervisor: Dr. Thang Tran Ngoc
Recipient of Merit Scholarship for Academic Excellence (all semesters)

High School for Gifted Students, VNU University of Science (HSGS)

High-Quality Class

πŸŽ“ High distinction (Top 1%) CGPA: 9.5/10
Sep 2017 – Sep 2020
Hanoi, Vietnam

Selected Research Experiences

Research Intern

NUS Web IR / NLP Group (WING)

Supervisors: Dr. Do Xuan Long, Prof. Min-Yen Kan
Research Topic: Test-time Methods for Agents

Jul 2025 – Present
Singapore

AI Research Resident

VinAI Research (acquired by Qualcomm)

Supervisors: Dr. Toan Tran, Dr. Tung Pham, Prof. Tan Nguyen
Research Topic: Efficiency, Scalability, and Robustness of Foundation Models

Mar 2024 – Present
Hanoi, Vietnam

Research Student

Modeling, Simulation and Optimization in Data Science (HATLAB), HUST

Supervisor: Dr. Thang Tran Ngoc
Research Topic: Multi-objective Optimization in Machine Learning

Aug 2022 – Aug 2024
Hanoi, Vietnam

Selected Publications

(*) denotes equal contribution

1. Artificial Intelligence in Metal Additive Manufacturing: Current Status, Challenges, and Future Developments.
Duc-Anh Nguyen*, Bui Truong Giang Le*, Van Anh Nguyen, Minh Tuan Vu, Manh Ha Bui, Minh Thanh Le, Trong Duc Nguyen, Tien Dung Hoang, Xuan Hai Le
Journal of Intelligent Manufacturing (Q1), Impact Factor: 7.4, 2025. link

2. Selective Sinkhorn Routing for Improved Sparse Mixture of Experts.
Duc-Anh Nguyen*, Huu Binh Ta*, Nhuan Le Duc, Tan Minh Nguyen, Toan Tran
Approved for Filing U.S. Patent, Qualcomm, 2025  |  Under review, 2025 arXiv

3. A Steepest Gradient Method with Nonmonotone Adaptive Step-sizes for the Nonconvex Minimax and Multi-Objective Optimization Problems.
Duc-Anh Nguyen, Tran Ngoc Thang
Under review, 2025 arXiv

For a complete list of publications, please visit the Publications page.