I am an Assistant Professor in the Department of Software Systems and Cybersecurity (SSC) at Monash University, where I lead a research lab working on trustworthy AI agents and cybersecurity. I am also the Founder of TokenTorrent (苍源智能), a startup building security middleware and infrastructure for AI agents — with the mission of commercializing cutting-edge security research into production-grade products that make agents safe, auditable, and trustworthy by design.

Before joining Monash, I was a Pre-Tenure Full Professor at USTC (2021–2025) and briefly held a tenure-track position at the University at Buffalo (2021). Prior to academia, I spent nearly two years at Meta/Facebook as a Research Scientist in the Security Infrastructure team, where I built large-scale distributed fuzzing platforms for continuous security testing across product lines, and developed detection methods against malicious proxy traffic and inauthentic social engagement. Earlier in my career, I was a Senior Software Engineer at Baidu, designing and building user management systems for location-based services including group buying and food delivery platforms — an experience that gave me a first-hand understanding of building internet-scale services.

My research vision is to instill trust into AI and agentic systems and the digital environments they operate in. I work on systematically uncovering security risks in LLMs, AI agents, and emerging network infrastructures, and on building practical defenses that combine machine learning with cryptographic primitives. Representative topics include AI agent red teaming and safety auditing, security of decentralized network infrastructures, network proxies and anonymous communication, and privacy-preserving detection and prevention of online abuse.