Cyber security graduate from the University of Edinburgh bridging AI security research and information security governance. Building compliance frameworks for clinical AI systems, conducting adversarial evaluations of frontier AI models, and engineering detection tooling across SIEM/XDR and clinical AI environments.
I am a Cyber Security MSc graduate from the University of Edinburgh with a foundation in Electronics and Communication Engineering from Anna University. My work spans two tracks: AI security research and information security governance.
On the research side, I build automated evaluation pipelines in Python for frontier AI systems (GPT-4, Claude, Gemini), run adversarial testing workflows, and develop detection tooling across SIEM/XDR stacks. As an AI Training Consultant at iMerit Technology, I evaluate model outputs against OWASP LLM Top 10 and MITRE ATLAS. As an Independent Security Researcher, I contributed to the Nemesys Insights adversarial evaluation of AI agents in ICS/OT environments for the AI Safety Fund / Frontier Model Forum.
On the governance side, I author risk registers, control mappings, and compliance documentation aligned to NIST AI RMF, ISO 27001:2022, and EU AI Act. I am building ClinRAG-Guard, an open-source compliance framework for clinical RAG knowledge base integrity with provenance tracking and retraction checking mapped to Article 10.
I delivered a conference talk at the Digital Research Conference 2026 at the Edinburgh Futures Institute on Zero-Knowledge Governance, proposing a cryptographic framework for auditing research integrity in Secure Multi-Party Computation environments.
Open-source compliance framework for clinical RAG knowledge base integrity. Implements provenance tracking, temporal validation, retraction checking against NICE, BNF, and NHS England. Mapped to EU AI Act Article 10 and MHRA AI regulatory framework. Targeting WISTP 2027 publication.
Adversarial evaluation of AI agents in industrial control systems for the AI Safety Fund / Frontier Model Forum. Assessed goal misalignment, data leakage, and privilege escalation risks in ICS/OT environments. Mapped findings to MITRE ATLAS.
Microsoft Sentinel SIEM with custom KQL detection rules for brute force, suspicious sign-ins, privilege escalation, and lateral movement. Full incident reports and operational runbooks mapped to MITRE ATT&CK.
Audit-grade ISMS portfolio for a fictional SaaS scenario. Full Annex A documentation, risk register, Statement of Applicability, internal audit execution with CAPA lifecycle and closure evidence.
Structured assessment of DVWA and OWASP Juice Shop across OWASP Top 10 (2021). 10 confirmed findings including SQLi (9.9 CVSS), file upload RCE, broken authentication, XSS, IDOR, and CSRF.
Supervised ML system classifying Tor Browser traffic with 81 to 88% accuracy using Random Forest, SVM, and XGBoost. Addresses anonymisation network detection as a threat to network visibility.
Governance framework combining SMPC and zero-knowledge proofs to audit research integrity without exposing sensitive data. Presented as a conference talk at Edinburgh Futures Institute, February 2026.
Classification system detecting phishing URLs based on domain identity, URL structure, and SSL indicators. Flask web app deployed on IBM Watson Studio. Collaborative undergraduate project (AICTE / IBM).
Edinburgh Futures Institute, University of Edinburgh, 26 February 2026. Presented a lightning talk on Zero-Knowledge Governance: Auditing Research Integrity in Secure Multi-Party Computation (SMPC). Session chaired by Professor John Vines on Ethics, Security and Integrity in Digital Research.
MITRE ATT&CK Foundations (AttackIQ)
🤖AI for Cybersecurity (LinkedIn Learning)
☁️AWS Academy Cloud Foundations
The Edinburgh Award (Students as Change Agents)
Tata Group Cybersecurity Analyst
Mastercard Cybersecurity
