Governed Engineering Traceability with Python and SQL

Governed engineering traceability with SQL, graph, and semantic retrieval

The problem Engineering impact reviews need to distinguish known relationships from plausible matches. A graph path can support a factual impact finding; a semantic match can surface a gap but still needs review.

The work I built a public Python demonstration using synthetic engineering data. SQL holds authoritative records, graph traversal retains the path behind each impact result, and semantic retrieval proposes additional candidates. The governance layer keeps those candidates separate, scores confidence and completeness, and requires human review where the evidence is incomplete. The repository includes tests and a local reference backend.

What this demonstrates Hands-on Python and SQL, data-quality checks, traceable analysis, and clear boundaries around AI-generated explanations. This is a portfolio demonstration, not a client deployment.