Supervision
Three PhDs and one MPhil completed, four PhD researchers in progress, and more than 40 BSc and MSc projects supervised at Cardiff since 2019, most of them applying machine learning to real attacks.
Continuous, AI-assisted compliance auditing for IoT security
Closing the gap between point-in-time audits and live IoT networks. The work combines human-centred continuous audit, knowledge graphs and retrieval-augmented LLMs for packet and log analysis, and ML intrusion detection for smart buildings and vehicle networks.
Key papers: Computers & Security 2026 · Journal of Cybersecurity 2026 · JCP 2026
Generative AI in cyber security education
Exploring how generative AI can be brought safely into security teaching, addressing transparency, accuracy and security, and using LLMs to help students analyse code vulnerabilities.
Key papers: Springer LNNS 2025 · CISSE 2026
Forecasting vehicle emissions trajectories using data-driven approaches
Developing data-driven methods and surrogate models to forecast how vehicle emissions will evolve, drawing on machine learning, digital twins and edge computing for sustainable, smart transport. Nima is also a Research Software Engineer at Cardiff, working on the security of EV charging infrastructure.
Related papers: ACM Computing Surveys 2025 · IET ITS 2026 · Profile
Operational technology security incident response
Researching how organisations detect, respond to and recover from cyber incidents in operational technology: the industrial control systems behind critical national infrastructure, studied through testbeds, cyber ranges and digital twins. Jenny also works on how to teach cyber-physical security, teaches at the National Software Academy and is a member of the NCSC ICS Community of Interest.
Key papers: ICERI 2025 · Profile
Intrusion detection for in-vehicle networks
Built intrusion detection for the CAN bus in modern cars. The thesis combined hierarchical federated learning, so vehicles learn from each other without sharing raw data, with a multi-stage detector for known and unknown attacks and a lightweight feature-selection method for CAN data.
Key papers: Vehicular Communications 2024 · Future Internet 2024 · Computer Networks 2026
Adversarial attacks on intrusion detection in connected and autonomous vehicles
Showed how attackers can fool machine-learning intrusion detection systems, including crafting traffic that triggers floods of false alarms, and what that means for defending connected and autonomous vehicles and IoT from insiders.
Key papers: Sensors 2024 · IEEE CSR 2024 · JISA 2022
Abuse of legitimate cloud services for command-and-control
Investigated how malware hides its command-and-control traffic inside trusted cloud and public services, and developed ways to detect it with dynamic malware analysis and machine learning.
Key papers: JCP 2023 (review) · JCP 2023 · IEEE ISNCC 2024
Exploring adaptive machine learning applications in edge IDS for vehicular systems
Investigated adaptive machine-learning models for intrusion detection that run at the edge, on or near the vehicle, so connected vehicles can detect attacks in real time as threats change.
Final-year BSc and MSc dissertation projects I have supervised since 2019, grouped by theme.
I welcome proposals in cybercrime and phishing, intrusion detection, adversarial and explainable machine learning, LLMs for security, and IoT or vehicle security. Please send a short proposal covering the scope, the state of the art and its limitations, your proposed methods, and the contribution you expect to make, to [email protected].