Wei Dong
I am currently a Temasek-Nanyang Assistant Professor starting Fall 2024. Previously,
I was a post-doctoral fellow at
Carnegie Mellon University,
co-supervised by
Prof. Elaine Shi and
Prof. Giulia Fanti.
Before that, I received my PhD from the
CSE Department of
HKUST, advised by
Prof. Ke Yi.
Our group has openings for Postdoctoral Researchers in differential privacy and cryptography. We are also always looking for Visiting Interns to join the group. If you are interested, please email me your CV.
Like many researchers, I believe one of the most important roles of a supervisor is to teach students how to become independent researchers. Beyond completing projects and publishing papers, I hope to pass on three skills that will benefit my students throughout their research careers.
-
The first is academic writing. I spend a considerable amount of time working with my students on their writing, often through several rounds of revision. Although this can take more time than rewriting the text myself, I believe it is one of the most valuable skills I can pass on. I am deeply grateful to my mentors, Ke, Elaine, and Giulia, who spent tremendous time helping me improve my own academic writing in my early age.
-
The second is academic reasoning and presentation: how to formulate a problem precisely, explain an idea clearly, and present technical results efficiently. A good result is not enough; we should also make it easy for others to understand why it matters and how it works. I am particularly grateful to Ke and Prof. Yufei Tao for teaching me a great deal about this.
-
The third is how to identify valuable research problems. A good problem should be both hard and important. “Hard” means that prior work or straightforward adaptations cannot easily solve it, while “important” means that it addresses a fundamental challenge rather than an artificial one. I would like to thank Ke and Prof. Xiaofeng Wang for greatly shaping how I think about research problems.
Research Interests
My research interests include data privacy and security, privacy-preserving AI, and LLM&AI security.
-
Data Privacy and Security
We design differential-privacy mechanisms for a wide range of data-analytics
tasks, aiming to safeguard privacy in analytical outcomes. Our focus is on
solutions that are both theoretically sound and empirically effective.
-
Privacy-Preserving AI
We leverage privacy-enhancing technologies such as DP, secure MPC, and ZKP to protect data privacy across the entire AI lifecycle—from data collection and training to deployment and inference.
-
LLM and AI Security
We study security and misuse challenges associated with large language models,
including preserving privacy in LLM applications, protecting LLM intellectual
property, defending against malicious uses, and using LLMs to enhance system security.
Selected Awards
- SIGMOD Jim Gray Doctoral Dissertation Award Runner-up 2024
- Best PhD Dissertation Award 2023, CSE, HKUST
- SENG PhD Research Excellence Award 2023, HKUST
- ACM SIGMOD Research Highlight Award 2023
- ACM SIGMOD Best Paper Award 2022
- Hong Kong PhD Fellowship 2018–2022
Service
- PC Member: ICDE 2024, Usenix Security 2025, VLDB 2026,
Usenix Security 2026, SIGMOD 2026, SIGMOD 2027, ICDE 2027,
Usenix Security 2027, EDBT 2027
Teaching
- SC2001 – Algorithm Design & Analysis: 2025/26 Spring
- SD6123 – Data Privacy in Data Science: 2025/26 Spring, 2024/25 Spring