Joshua Biondo · Selected work · Current role · Data platforms
Metadata-Driven Lakehouse
One pipeline, every source. New tables onboard with a config row.
Outcome: 15+ hours a week of manual reporting eliminated ($35K+ a year)
What I built
- A modular medallion lakehouse ingesting millions of rows from APIs and AWS S3, with credentials in Key Vault.
- One metadata-driven pipeline across 6+ sources and 50+ tables.
- Schema contracts in Silver that fail the run when a source drifts.
Stack: Fabric, PySpark, Delta, Key Vault