At University of Stirling, I built a reproducible pipeline for extracting IoT-23 and Gotham 2025 network flows from packet captures, then trained and evaluated LSTM, GRU, and XGBoost intrusion detectors. My MSc research examined how changing network environments affected detection and compared seven adaptation strategies.
I found that detection rates dropped sharply for several attack types after environmental changes, and traced the failure mainly to miscalibrated confidence. Drift-triggered retraining recovered performance using 54–66% of the labelled data needed for scheduled retraining.
For my Glasgow Caledonian University honours project, I built a real-time hybrid malware detector combining XGBoost for known attacks with Isolation Forest for anomalies. It analysed live network traffic and raised JSON alerts with SHAP feature attributions.
As a Volunteer AV and Network Technician at RCCG City of God Glasgow, I designed an audio-over-IP network with an external AV integrator and configured switches for audio traffic. I also manage wireless microphone RF and run live sound and AV for services and events.

