Hackathon prototype
GRC AI.
I led a team building a local-AI prototype for reviewing policies against compliance frameworks.
Why I made it
Reviewing policies against several compliance frameworks involves a lot of reading, cross-referencing, and handling documents that may be sensitive.
What I built
I led a team building a prototype that uses local AI to analyze policy documents against multiple compliance frameworks. We used Ollama and Gemma 2 2B to explore the workflow without relying on a hosted model.
What I learned
The useful question was where AI could support a review and where a person still needed to make the decision. A model's answer is a starting point, not proof of compliance.
