video1.0<iframe src="https://www.loom.com/embed/a776e6f70c7540c399f30bb72140058d" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/a776e6f70c7540c399f30bb72140058d-a490743263ebfc24.gif253.526This Loom demonstrates an agent that can generate production data pipeline code while validating it against governance policies and the existing data schema. The prototype uses FastAPI for the backend and ReactJS on the frontend, where a user requests a pipeline to transform raw customer transactions from a data hub schema into a daily customer revenue model. After fetching all relevant data schema, the system creates an MCP context, sends it to the Gemini API to generate artifacts including staging SQL, DPD models, DBT schema, Airflow workflows, and a README and test cases. It then performs validation checks such as column and governance rules, documents handling and validation rules, and creates and merges a pull request to the project main branch once the PR is ready.