Every problem is deliberately vague. Your prompts get scored, not the AI’s answers.
The app needs to notify users.
Users need to send messages to each other in real time, with support for one-on-one and group conversations.
Store and serve arbitrary objects (images, backups, datasets) with eleven-nines durability, strong consistency, and exabyte-scale capacity — similar to S3.
Engineers push code changes that must be built, tested, and rolled out to thousands of servers with support for canary releases and instant rollback.
Multiple users edit the same document simultaneously and see each other's changes in real time, similar to Google Docs.
User-generated content must be screened for policy violations before or shortly after publishing. Design for scale, accuracy, and human review workflows.
Track how many times each piece of content has been viewed. The counter must be fast to increment and accurate enough for analytics.
Customers order goods for same-day or on-demand delivery. Design the order management, driver dispatch, routing, and real-time tracking systems.
Services are hammering the database. Design a distributed caching layer (like Redis) to absorb read traffic and reduce latency.
Multiple services need to coordinate access to shared resources. Design a distributed lock manager that is correct under network failures and node crashes.
Services need to communicate asynchronously at scale. Design a durable, ordered message queue that guarantees at-least-once delivery.
Build a globally distributed relational database that supports ACID transactions, horizontal scaling, and multi-region replication with tunable consistency.
In a microservices architecture, engineers need to trace a request across dozens of services, correlate spans, and identify bottlenecks — at millions of requests per second.
Multiple services must participate in a single transaction that either fully commits or fully rolls back — even across failures and network partitions.
Clients resolve domain names to IP addresses. Design a hierarchical, globally distributed DNS that handles billions of queries per day with low latency.
Engineers need to toggle features on or off for specific users or percentages of traffic without redeploying code.
Users need to store, sync, and share files across devices — similar to Dropbox or Google Drive. Design for reliability, deduplication, and large-file support.
A limited quantity of items goes on sale at a fixed time. Millions of users attempt to purchase simultaneously. Design for fairness and correctness under extreme load.
Customers browse restaurant menus, place orders, and track delivery in real time. Design the order lifecycle from placement to doorstep.
Flag suspicious transactions in real time before they complete. Design the feature pipeline, model serving layer, and decision engine at payment scale.
