The Baldwin Group · 2026

Unstructured document automation

Databricks pipelines that select, retrieve and deliver unstructured policy documents by metadata.

Databricks · Python · Unstructured data

Problem

Policy documents were indexed only by metadata, and downstream teams pulled them by hand.

Work

  • Pipelines that select documents by metadata and keywords, retrieve them and deliver approved files downstream.
  • Primary engineer: 330+ commits and 50+ pull requests in five months.
  • Project instructions checked in for AI coding assistants.

Result

Scheduled, auditable retrieval.

← All work