Om Shah

Research paper on weaponised AI that can select and engage targets with reduced human oversight. Uses open-source case studies (e.g. SGR-A1, Super aEgis II, Guardium-LS) to discuss accountability, international law, sensor and recognition limits, dataset bias, and options for oversight and regulation. Written for educational and research purposes only.

Written assessment on cloud governance practices and structured threat analysis for cloud workloads, including control trade-offs and risk framing relevant to organisational security posture.

Same UTS capstone as Sentinel: a full project development journal covering the virtual lab (Kali, Ubuntu, Windows Server on VMware/VirtualBox), recon and spoofing labs, early Python work on Sentinel, and weekly reflection on what worked, what failed, and what was learned.

Compares Naive Bayes, SVM, and a deep learning 1D CNN for spam detection, with tokenisation, padding, embeddings, training behaviour, and evaluation through precision, recall, F1, confusion matrices, and ROC-AUC. The CNN delivered the strongest overall performance while balancing detection with false positives.

Ethics and professional practice analysis for a resort guest hospitality system: data collection and retention, RFID location tracking, and childcare safety checks, grounded in ACS, PMI, and IEEE codes of ethics, with practical recommendations to protect privacy, safety, and public trust.