Final-year BSc Cybersecurity at Liverpool John Moores University, ranked top 3 of 62. Predicted First Class Honours. KLCP expected June 2026. Building tools at the intersection of AI and offensive security.
Automated threat intelligence aggregation tool monitoring clearnet sources for emerging CVEs and zero-day disclosures. Scrapes Reddit, GitHub, NVD, and CISA KEV. Uses Claude Haiku to parse unstructured text into structured CVE records with severity, exploit status, and affected software classification. Deployed on a live VPS with nginx, Docker, and HTTPS.
Novel BERT-based classifier for Windows API call sequences using a custom multi-granularity direct integer encoding scheme, bypassing standard WordPiece tokenisation. Achieved 97.3% F1-score on MalbehavD-V1, exceeding state-of-the-art baselines. Demonstrated cross-dataset transferability on MAL-API-2019 with shot-learning adaptation.
Group project investigating LLM jailbreaking techniques via prompt injection and AI-generated malware. Developed defensive countermeasures and presented findings to faculty.
Built production-grade web applications following OWASP Top 10 security standards, implementing authentication, parameterised queries, XSS/CSRF protection, and AES encryption. Achieved 88% module grade.
// Available for graduate roles from September 2026. Open to cybersecurity operations, penetration testing, and security consultancy.