The problem

Organizations with multiple locations run different systems (EHR, HR, finance). Each dumps raw data in different formats and schemas. We want to unify this into a central semantic model – without forcing locations to change their systems.

What we’re trying to achieve

• Automatically understand what’s in a raw database schema
• Detect when tables across different sources represent the same entity
• Merge and map local data to a central ontology
• Make this process automated with human‑in‑the‑loop supervision

One route we’re exploring

Cube‑based semantic layer(s) with agent‑assisted schema generation and entity matching. But we’re open to other approaches: dbt, custom tooling, graph‑based solutions, etc.

Your profile

• 5+ years data engineering
• Deep experience with data modeling and schema design
• Worked with multi‑source data integration challenges...