Build a Distributed Session Store for 50M Active Sessions Premium
Design a session storage system that maintains 50 million active sessions with sliding expiry, serves session data in under 5ms, and handles node failure without logging users out.
Build an LLM Prompt Caching Layer Premium
Design a semantic caching layer for LLM API calls that identifies semantically equivalent prompts, serves cached responses for near-duplicate queries, and reduces inference costs without degrading response quality.
Build Netflix's Content Delivery Network Premium
Design a CDN that caches the right content at the right PoP before users request it, handles cache misses without hitting origin, and evicts stale content across 200+ edge locations globally.
Build a Real-Time ML Feature Store Premium
Design a feature store that serves pre-computed features for online ML inference in under 5ms, keeps features consistent between training and serving environments, and supports point-in-time correct feature lookups for offline training.
Build Reddit's Post Ranking Algorithm Premium
Design a post ranking system that combines upvotes, downvotes, time decay, and engagement signals into a score that keeps feeds fresh without requiring full recomputation on every page load.
Build TinyURL - A URL Shortener at Scale Premium
Design a URL shortening service that generates collision-free short codes, handles 100,000 redirects per second with sub-10ms latency, and tracks click analytics without slowing down redirects.