PFZ LLM Auditor Agent
This system is built as a Flask application and is deployed on Google Cloud Run, leveraging the Vertex AI SDK to interact with Google's foundation models. This ensures high-quality, trustworthy outputs and provides a robust foundation for building responsible AI applications.
The agent is containerized using Docker, and its deployment is automated via Cloud Build, which is triggered by pushes to a Git repository. The system has been successfully deployed and is now fully operational on Cloud Run, with the core logic encapsulated within a single, highly-engineered prompt.
The PFZ LLM Auditor Agent is a powerful, multi-capability system designed to evaluate and refine AI-generated answers for factual accuracy and trustworthiness. It operates in a two-step process: first, it fact-checks an LLM's answer by finding supporting evidence, and then it produces a revised and more accurate answer.
Our flagship Enterprise AI Agent