Crack the data engineering system design round with a 6-step framework and sizing cheat sheet. 8 fully worked designs: Postgres CDC to a lakehouse, a 400M-user segment builder, real-time ad targeting, partner reporting, a bitemporal security master, sessionisation, point-in-time features, and a cost-optimised Spark on EMR platform with Airflow. Each covers requirements, numbers, architecture diagram, trade-offs, data model, failures and backfills, data quality, cost, and interviewer follow-ups. 71 pages, 36 diagrams.