Do you work on my team? This is exactly how we're using Duckdb with Databricks as the massive data bulldozer and Duckdb as the scalpel.