I work on runtime defense for AI agent systems — prompt-injection detection, policy enforcement at the tool-call boundary, and governance for multi-agent meshes.
Auren — an AI-agent defense platform, in pre-GA development.
The code is private while it's in beta.