Junyi Hou (侯君宜)
Junyi Hou is a PhD student in the Systems & Networking Research Lab at the National University of Singapore (NUS), supervised by Prof. Bingsheng He. Previously, he collaborated with Dr. Zhaomin Wu. He earned a Master of Computing in Computer Science from NUS. He is passionate about programming and deeply immersed in exploring innovative solutions. His research interests focus on Infrastructure and Systems for Large Language Model.
news
| Jun 10, 2026 | 🎉 PaperDebugger — our “Cursor for Academia” — is accepted to the WWW 2026 Demo Track! Check out the code. |
|---|---|
| Jun 01, 2026 | I received a Google TPU Builders Award, including Google Cloud credits and access to flex-start TPUs, to support my work on TPU-based LLM inference and backend development. |
| May 01, 2026 | Our project, Lightweight and Automated Performance Optimization of Training and Inference on TPUs, led by Prof. Bingsheng He, was selected for the 2026 Google Awards for Machine Learning Research and Education with TPUs. I am contributing to the project as a PhD researcher. |
| Oct 15, 2025 | 🏁 Serving as the Web Chair for ICPP 2026. Submission details are available in the Call for Papers. |
| Sep 25, 2024 | 📄 One paper is accepted by NeurIPS 2024 |
education
| 2025 - Now | National University of Singapore PhD Candidate |
|---|---|
| 2022 - 2024 | National University of Singapore Master of Computing in Computer Science Specialization |
| 2017 - 2021 | Macau University of Science and Technology Bachelor of Science in Software Technology and Application 🥇 GPA: 3.92/4.00, Rank: 1/150 |
work experience
| 2024 - Present | National University of Singapore Research Assistant in the Department of Computer Science |
|---|---|
| 2021 - 2022 | Institute of Automation, Chinese Academy of Sciences (Beijing) Research Intern in the National Laboratory of Pattern Recognition |
| 2020 | Tencent Technology (Shenzhen) Back-end Developer Intern, Cloud Arch & Platform Dept. 🥇 1st Prize in the Internship Project Competition |
publications
- Dual Balanced Class-Incremental Learning With im-Softmax and Angular RectificationIEEE Transactions on Neural Networks and Learning Systems, 2024
- VLDB 2024
LLM-PBE: Assessing Data Privacy in Large Language ModelsIn 50th International Conference on Very Large Data Bases, VLDB 2024, Guangzhou, China, August 26-30, 2024. Our LLM Privacy Challenge can be found here , 2024 - NeurIPS 2024
Federated Transformer: Scalable Vertical Federated Learning on Practical Fuzzily Linked DataIn The Thirty-eighth Annual Conference on Neural Information Processing Systems, NeurIPS 2024, Vancouver, Canada, Dec 9-15 , 2024