Designing Data-Intensive Applications, Second Edition

This book by Martin Kleppmann and Chris Riccomini explains how systems that store and process data really work, from indexes inside a single database to consensus across dozens of nodes. Rather than surveying specific products, it examines the principles they are built on, so after reading it the documentation for PostgreSQL, Kafka, or Cassandra looks completely different. The second edition is updated for the modern stack: cloud services, data warehouses, and data lakes.

It is best read in order, since each chapter builds on the previous ones. The book is not about interviews, but after it, system design questions stop being a set of memorized answers.