Making AI operational in constrained public sector environments

Build a secure-by-default, offline-capable inference stack optimized for public sector environments using tiny language models. The stack must enforce strict data residency and logging requirements.

Suggested repo: gov-ai-stack

"High-security local inference for sensitive environments."

Estimated effort: 100h