2026 · Diploma Project · International Information Technology University (IITU)
An interactive AI-driven cybersecurity simulator that deploys adversarial AI agents to stress-test a simulated smart-home IoT environment. The system evaluates LLM responses to prompt injection, context manipulation, and privilege escalation attacks — providing real-time risk scoring and analytics.
Project lead and main developer. Built Imperium AI — an interactive AI Red-Teaming Battle Simulator for smart-home security research. Specialises in AI systems, cybersecurity research, and interactive 3D visualisation.
Researches attack scenarios and agent strategies for the project. Focus: improving scenario quality, expanding the Learning Journey content, and making attack visualisation more practical. Specialises in prompt injection and context-poisoning research.
Responsible for 3D scenes, IoT object models and interactive prototype pages. Works on raising the realism of the smart-home visuals on the battle / view pages, and on a voice-driven IoT control prototype.