Build an AI Agent Task Queue with Resumability Premium
Design an agent execution framework that handles long-running tool calls, persists agent state so tasks can be paused and resumed across server restarts, enforces token and cost caps, and supports parallel sub-agent execution.
Build a Blue-Green Deployment System Premium
Design a deployment system that can switch 100% of production traffic from the old version to the new version in under 30 seconds and roll back instantly if error rates spike.
Build a Chaos Engineering Platform Premium
Design a controlled failure injection platform that safely introduces latency, packet loss, and resource exhaustion into production services, enforces blast radius limits, and automatically halts experiments when SLOs degrade.
Build a CI/CD Pipeline Orchestrator Premium
Design a system that handles 10,000 concurrent build and test jobs, assigns them to workers, streams logs in real-time, and ensures no job is lost even if a worker crashes mid-run.
Build a Concurrent Device Session Manager Premium
Design a session enforcement system that limits streaming to N concurrent devices per subscription plan, kicks the oldest session when the limit is exceeded, and works correctly across distributed servers.
Build a Distributed Cron Job Scheduler Premium
Design a cron scheduler that guarantees exactly-once execution of jobs across a cluster of nodes, handles missed executions during downtime, and scales to millions of scheduled tasks.
Build a Distributed Key-Value Store Premium
Design a distributed KV store that partitions data using consistent hashing, replicates each key across N nodes for durability, and handles node failures with tunable consistency guarantees.
Build a Distributed Lock Manager and Leader Election Service Premium
Design a distributed coordination service that provides mutual exclusion across a cluster, implements lease-based locks that expire on node failure, and enables leader election for singleton workloads.
Build a Distributed Unique ID Generator Free
Design a system that generates globally unique, roughly time-sortable IDs across thousands of nodes with no coordination overhead and zero collisions under any failure scenario.
Build a Distributed Write-Ahead Log Premium
Design a durable WAL that survives single-node crashes, replicates log entries across nodes before acknowledging writes, and supports point-in-time recovery by replaying from any checkpoint.