Practice briefStart with your own approach
28. Aggregate events that arrive late or twice
Keep tenant usage totals correct across duplicate delivery, late events, outages and replay.
Data engineeringMedium
Guided practice
Start with a template. Work through each step. Ask Coach when you need a second opinion.
All designs 8
Ordered by design numberCompany tags are community-reported. Counts on cards show how many people reported that design.
StreamingClear filters
40. Design an Ad Click Aggregator
Count every ad click once, fast enough to chart live and exactly enough to bill. One-hour interview boards for junior, senior and staff: requirements, data layer, low-level design and what goes wrong at every component.
ClassicMedium41. Ad Click Aggregator: Low-Level Design
The classes and code behind the ad click aggregator: signed click tokens, a redirect that never waits on the log, and a stream counter that de-duplicates and handles late clicks. Tests run all of it.
Data engineeringMedium43. Design YouTube Top-K
The K most-viewed videos for the last hour, day, month and all time from 700,000 views a second, exactly and in milliseconds.
ClassicMedium72. Design a Distributed Log like Kafka
A partitioned, replicated append-only log: a million messages a second, order per key, nothing acknowledged ever lost, a week of replayable history.
ClassicMedium77. Design a Stream Processor like Flink
A stateful stream processor: dataflow graphs, keyed state in RocksDB, event time and watermarks, windows, barrier checkpoints, exactly-once into Kafka.
ClassicMedium85. Design a Change Data Capture Pipeline
Every committed change in Aurora and DynamoDB, read from the log and delivered in order per row to search, caches, Redshift and an S3 lake.
ClassicMedium94. Design a Driver Heatmap
Count 4 million moving drivers per map cell from a million pings a second, and serve the density to a million viewers as cached map tiles.
ClassicMedium