Doctoral role in AI security exploring information-theoretic foundations of memory poisoning for autonomous agents. Combines information theory, statistical learning, communication theory and LLM experiments to develop rigorous attacks, defenses and detection limits.