How would you design a data pipeline to ingest, clean, and standardize millions of heterogeneous insurance claims records for Amino?
