Postgraduate Researcher – UWE (2022 – Present)
Developing AI/ML models for fraud and anomaly detection in telecom systems. Presented at conferences and published in peer-reviewed venues.
Email - james6.barrett@uwe.ac.uk
LinkedIn | GitHub | Scholar | LeetCode | UWE Profile
Developing AI/ML models for fraud and anomaly detection in telecom systems. Presented at conferences and published in peer-reviewed venues.
Integrated XAI/ML into real-world telecom infrastructure with a focus on product-oriented cybersecurity enhancements.
Lecturing on advanced cybersecurity and ML vulnerabilities while mentoring student research projects.
Led modules in cybersecurity and computing; supported student research and practical coursework.
Upgraded backend systems in PHP/Linux and deployed secure applications using Git workflows.
Focus: Fraud detection, anomaly mitigation in telecoms, real-world XAI applications.
Dissertation on maritime network frameworks for security and mitigation.
Access control and login form frontend/backend with PHP and MySQL.
Full-stack DB/website system for managing registered information (FdSc dissertation).
Portable base station with cellular interception functionality (BSc dissertation).
Pen-testing toolkit tailored for maritime protocols (MSc dissertation).
Weighted time series forecasting models with SHAP interpretation.
Tool for generating telecom datasets for research simulation.
Interactive XAI viewer for time series data (Streamlit + SHAP).
Collection of websites built during web development course.
Collection of proof-of-concept malware and DDoS tools.
Challenges in using CDRs with ML for analytics (paper).
Forecasting analytics for telecom maintenance (paper).
Simulation-based time series dataset generation project.
Experiments uncovering root cause analysis in real-world scenarios.
ML temporal method included in official patent (Ribbon Communications).
Planned publication on novel XAI temporal method.
Early research on network time series data aggregation.
Exploring how causes change over time in RCA (early paper).
Sankey, box, and custom visual methods for SHAP and LIME.
© James Barrett