Since 2009, I have been developing autonomous and provable techniques for securing the information environment. This means building and configuring systems that are secure by design, not just addressing problems when they strike. By leveraging tools from graph theory, game theory and AI/ML we build practical solutions that help human operators deal with the complexity, fast-paced and deceptive nature of the cyber environments. Our research program is supported by multiple grants from Defence and the Australian Research Council.
Current research themes
- AI agent security. Memory, document, skill and tool poisoning, agent misbehaviour, alignment, action interpretation, and secure agent architectures.
- Autonomous enterprise defence. Active Directory attack graphs, attack-path management, decoys, honeypots, and robust network hardening.
- Cyber-physical and OT security. Edge AI, CAN and CAN FD security, OT/IT boundary assurance, and resilient operational platforms.
Selected completed projects
- Resilient Platform Interfaces, completed 2026. Defence Trailblazer collaboration with Cisco and Data#3. The project developed and demonstrated methods for protecting the OT/IT boundary using a simulated vehicle environment, CAN bus interfaces, a Cisco industrial router, and edge-AI models.
Research translation
Our research translation activities produce demonstrators, software, datasets, and operational prototypes that allow research outcomes to be evaluated with industry and government partners.
