Mayur Sinha

About Me

I am a Staff Security Engineer focused on AI and cloud-native security. I build practical security systems across application security, infrastructure security, and detection engineering.

My core focus is AI security: securing LLM and agent workflows, stress-testing systems for prompt injection and data exfiltration risks, and building guardrails teams can run in production.

Over the last 15 years, I have worked across startups and enterprise environments in research, product security, and incident response, from malware reverse engineering to modern DevSecOps programs.

I also collaborate with a non-profit on long-horizon AI risk research and write about AI security, engineering systems, and the deeper philosophy of intelligence and safety.

If you are building secure AI systems or exploring similar questions, I would love to connect.

Featured Projects

All projects

MCP security evaluator
Tooling to evaluate and harden MCP and LLM integrations.

Builder2025
Adversarial ML security lab
Hands-on lab for adversarial ML attack and defense experiments.

Builder2025
BERT-based RASP for runtime security
BERT-enhanced runtime application security for in-app threat detection.

Researcher2023