AI/ML PENTEST // VA/PT // RED TEAM // CLOUD SECURITY
— Security Researcher
I help organisations stay one breach ahead — building offensive simulations, hardening cloud estates and running 24/7 SOC operations that actually catch real threats.
ID // 1024-SUNIL-T // CLEARANCE: RED
Over years of hands-on work across penetration testing, incident response and cloud security, I've shipped findings that moved security postures from compliant-on-paper to genuinely resilient. My approach is part adversary, part architect: break it loud, fix it quietly, document everything.
I work across pentest, SOC and cloud — bridging offensive findings with defensive engineering, and documenting work that survives team turnover.
Adversarial assessment of a production LLM platform. Bypassed guardrails via indirect prompt injection through RAG sources, achieved data exfil from connected tools, and reported model-extraction risk.
Full-scope adversary simulation against a Tier-1 bank: phishing, OT pivot, domain escalation. Closed with C-suite tabletop debrief.
Reviewed 6 production AWS accounts. Found IAM privilege creep, exposed S3 sinks and crypto-mining via stale Lambda. Hardened landing zone.
Authored Falco + Splunk detections for container escape, sidecar tampering and crypto-jacking. Now part of org-wide baseline.
For engagements, advisory work, talks or just to trade war stories — send a message. Most replies within 48h.