Public skill

data-throughput-accelerator

Diagnose and accelerate large data movement — ingestion, backfill, export, ETL, warehouse loading, manifest catch-up, and table synchronization — by isolating the true bottleneck, benchmarking variants, and codifying the fastest path with a hard accounting block proving rows and timestamps cohere. Use when a pipeline or backfill is too slow and must get faster without losing data correctness.