Design a customer warehouse with SCD Type 2
Keep a trustworthy history of how customer attributes change over time.
Data engineering interviews test whether you can move, model and trust data at scale. These rounds cover star-schema and slowly changing dimension warehouses, change data capture, real-time analytics and ad click aggregation, sessionization, IoT telemetry, log pipelines, lakehouses, reusable ETL and ELT frameworks, payments reconciliation, A/B test analytics, data quality and user data deletion.
You design the pipeline end to end, from sources and ingestion through storage, transformation and serving, and explain the guarantees you chose: batch or streaming, exactly-once or at-least-once, freshness versus cost.
Keep a trustworthy history of how customer attributes change over time.
Turn a large event stream into operational metrics people can act on.
Give restaurants live order metrics and executives daily reports from one stream of order events.
Give many teams a consistent way to build, run, and operate data pipelines.
Collect trustworthy product events from web and mobile clients.
Notify the right owner when pipelines, data freshness, or business metrics go wrong.
Join streaming payments with historical signals for fast, explainable fraud decisions.
Help teams detect, diagnose, and prevent incorrect or stale data.
Model orders, products and customers so analysts can answer sales questions quickly and correctly.
Stream every insert, update and delete from a production database into the warehouse within minutes.
Count ad clicks in near real time for advertiser dashboards, accurately enough to bill on.
Ingest readings from millions of devices and serve live and historical views of each one.