At Secvera Technologies, I oversee technical scoping and delivery for penetration testing, cloud security, and compliance engagements. I review findings and CVSS severity, and mentor peers and junior analysts on testing methodology and report writing.
As an independent cybersecurity researcher, I’ve provided penetration testing and compliance consulting to clients across fintech, SaaS, and healthcare. I’ve conducted manual web application tests, reported OWASP Top 10 vulnerabilities, and delivered CVSS-scored findings with remediation guidance.
My research focuses on AI security and the LLM supply chain. I found tokenizer and poisoned-RAG attack paths, including a malicious tokenizer that injects jailbreak prompts and a path from a web page to remote code execution on an MCP-enabled assistant’s host.
I also developed AgentFlowGuard to track data provenance, propagate taint, and control tool execution across agent workflows. In AI judge research, I built a confidence-calibrated judge with human routing and open-sourced three components with their methodology.

