Whenever you're ready, walk me through how you'd approach this.
Tell me about designing a data pipeline for a real production use case, including how you handle ingestion, transformations, and delivery.
Present one production example and cover source contracts, batch or streaming ingestion, raw and curated storage, transformation strategy, orchestration, idempotency, schema evolution, data quality, observability, backfills, and recovery. Explain the design trade-offs and how you would validate that delivery meets freshness, correctness, and availability expectations.