When multi-agent systems beat single-agent, the supervisor / worker / reviewer / critic patterns, hierarchical vs swarm, communication models, evaluation, and the production patterns that survive contact with users.
End-to-end system design for a WhatsApp / Telegram-style chat. WebSocket fleet, message storage, delivery receipts, online presence, group chat fanout, end-to-end encryption, and the operational realities of running it at scale.
End-to-end system design for a Twitter / News Feed service. Functional + non-functional requirements, capacity, schema, fanout-on-write vs fanout-on-read, hybrid for celebrities, ranking, caching, and what interviewers love.
Production patterns for idempotency keys, retry strategies, the outbox pattern, and the truth about exactly-once delivery. The patterns every backend engineer needs to handle network failure correctly.
When to pick Kafka vs NATS vs RabbitMQ. Semantics, throughput numbers, ordering and persistence guarantees, operational cost, and a decision tree based on what you’re actually building.
End-to-end system design for a URL shortener — requirements, capacity estimation, ID generation (counter, hash, base62), schema, caching, analytics, abuse prevention, and the parts most write-ups skip.
A backend engineer’s working mental model for distributed systems. Failure modes, CAP, replication, partitioning, consistency, time and ordering, idempotency — explained without the textbook.