Built three production event-driven pipelines on GCP to sync DV360 advertising data to AWS SQS.
Metadata: 8 entity types delivered in dependency order across 6 tiers using Cloud Tasks chaining and parallel Cloud Run Jobs.
Reports: auto-detected new DV360 report tables from BigQuery audit logs and normalized fields to a common schema.
ADH match-rate: 6-state async workflow for UPDM + match-rate queries with privacy cooldown handling.
All pipelines used outbox delivery with receipt-based reconciliation.