About Me
I have tried a few different paths, but research is still what I enjoy most.
As a child, I loved the satisfaction that came from solving mathematics problems. Looking back, it was one of the earliest signs of my love for problem-solving. My earliest research came through my computer science theses at Imperial (2014) and UCL (2018). During my PhD, from October 2019 to February 2025, I worked mainly on blockchain with my supervisor, Arthur Gervais, and my collaborator, Kaihua Qin. I was incredibly fortunate to have Arthur as my supervisor. Working with Arthur and Kaihua was when I truly began to take research seriously.
I am now at the University of Sydney, where I supervise students in their research and teach INFO5995. I reflected on the course in a LinkedIn article. I still try to stay hands-on and contribute directly to research.
I study the security and reliability of AI agents. My current work asks how agents can discover and validate real vulnerabilities, learn reusable knowledge through experience, and make decisions grounded in trustworthy evidence. More broadly, I am interested in agentic systems that can generate and validate new knowledge beyond security. I believe AI will fully replace researchers in the near future, and that the mission of our generation is to make it happen.
Links
Media coverage
• lzhou1110 (at) gmail (dot) com — use this for all contact except USyd-internal matters; Gmail is preferred because it works better with agent-assisted communication