My research spans two parts: hardware security and AI security.

In hardware security, I develop and analyse methods to uncover and mitigate vulnerabilities in pre-silicon and post-silicon haradware, including deep-learning enhanced side-channel analysis, hardware fuzzing, reinforcement-learning guided testing, and microcode-guided CPU evaluations.

In AI security, I investigate security and trustworthiness issues in modern artificial intelligence, including attacks and defenses on large language models and their deployment in safety-critical contexts, exploring both algorithmic and system-level threats.