Research Associate / Senior Research Associate / Research Fellow in Security Assurance for Agentic

The Smart Internet Lab at the University of Bristol invites applications for a researcher to join the AURA (Autonomous Unified RAN Architecture) project. The appointee will address critical security challenges in AI-native future networks, focusing on agentic AI security assurance, adversarial testing, and zero-trust policy enforcement. Responsibilities include identifying threat landscapes, simulating multi-agent attacks, and designing runtime decision validation engines. The role involves close experimental evaluation using partner platforms and university infrastructure, as well as publishing research in high-impact venues.

• Identify and document threat landscapes across data ingestion, intent interface, and execution layers of agentic AI systems.
• Simulate and demonstrate agent misuse and multi-agent coordination attack scenarios on live partner platforms.
• Design and implement intent-to-policy mapping to translate agent intent into verifiable formal policy constraints.
• Develop runtime decision validation engines to evaluate proposed configuration actions against cross-layer security dependency models.
• Implement dynamic trust scoring mechanisms based on historical agent behavior and continuous behavioral monitoring.
• Develop automated mitigation strategies including rollback mechanisms and isolation protocols for compromised agents.
• Publish research findings in high-impact journals and present at leading security conferences.
• Evaluate research experimentally using partner agentic AI platforms and JOINER national open-networks infrastructure.