System design interview question
Design a Distributed Real-Time Leaderboard and Rankings System like Twitch's Concurrent Viewer Counter and Esports Rankings at Scale
Design a system that maintains real-time global leaderboards across multiple dimensions (daily/weekly/monthly scores, regional rankings, category-based leaderboards) for a competitive gaming platform serving 500M users with 50M monthly active participants. The system must support 100K rank update events per second, sub-500ms query latency for top-100 leaderboards, handle geographic distribution across 6 regions, and maintain eventual consistency while preventing ranking manipulation through anomaly detection. Users should be able to query their rank, nearby competitors, and historical rank trends. Address concurrent writes, leaderboard tie-breaking rules, stale reads during peak load, and efficient recalculation of ranks as scores change.
Practice this question step by step with an AI interviewer that reviews your requirements, your architecture and your trade-offs, and scores the design.