Selected work

AI safety infrastructure

Agentic AI Security

A policy enforcement layer that inspects LLM traffic for prompt attacks, data leakage, and unsafe output.

Engineering outcome

Every model request inspected

Every inbound and outbound model request passes through a configurable safeguard layer.

Delivery

Security engineering

FastAPILlamaGuardDocker

Build something useful

Have a system that should work better?

Talk to an engineer