A focused track on the systems most often asked at top-tier interviews — payments, ride-hailing, rate limiting, and the distributed-systems primitives that underpin every answer.
The first layer every request touches. Get this wrong and the system never gets a chance.
Geo-spatial sharding, location streams, and matching at city scale.
Idempotency, fraud, settlement — when "eventually consistent" is not allowed.
The math behind every "it depends" answer in a system-design loop.
You can have at most two of consistency, availability, and partition tolerance — and partition isn't optional.
How to shard data across servers so adding or removing a node only moves a small fraction of keys.
How a cluster of nodes elects a leader and replicates a log so they stay in lockstep — without a coordinator.