A field guide to the data stack at the level of an IC5 / staff system-design interview. Storage internals, streaming semantics, lakehouse formats, and the operational craft — built around interactive simulators of every concept that's usually drawn on a whiteboard.
App backends, mobile clients, IoT sensors, third-party APIs. Every event has a creator.
An ordered, durable, partitioned log. Decouples producers from consumers. The cleanest abstraction in this whole stack.
Stream jobs filter, enrich, window, aggregate. Batch jobs do the same, just on bounded data.
Object storage holds the raw. Open table formats give it ACID. Indexes give it speed.
OLAP engines for BI, vector stores for ML, key-value stores for online features.
Dashboards, ML features, billing, fraud, the recommender. The stack only matters because of this row.
Linear the first time — each lesson assumes the last. Reference order after that. The capstone (lesson 12) is a 45-minute IC5 system-design walkthrough you can step through one move at a time.
Progress is tracked in your browser. No accounts, no servers. Reset anytime from the footer.
The stack, top to bottom — in twelve minutes. Then the rest of the course is just zooming in.