I study intelligent systems where models meet the world — agents, security, and markets.
PhD researcher at the University of Melbourne studying the security of LLM-based systems under constrained external interaction.
I am supervised by Xingliang Yuan (opens in a new tab), Bang Wu (opens in a new tab), and Shangqi Lai (opens in a new tab).
Security and privacy in language models.
Security and reliability in AI agents.
Security in systems with multiple interacting participants.
Published research and independent exploration.
Published researchSTMUS 2025 workshop
Can model behaviour reveal what was meant to be forgotten? I developed and evaluated a probe-based pipeline comparing original and unlearned LLMs to infer removed subjects.
Read the paper (opens in a new tab)Independent explorationIn development
An independent project exploring how AI agents could take a market question through a full research cycle.
About the ideaI studied Computing and Software Systems as an undergraduate and completed a master’s in Computer Science. My PhD now focuses on the security of LLM-based systems, from individual models to agents and multi-principal systems.
Much of what interests me lives at a boundary: between model and system, prediction and action, certainty and uncertainty. I like building around those boundaries — and then testing the assumptions that hold them together.