Cheng demonstrated solid understanding in core system design topics and foundational backend skills during our mock interview focused on Game Leaderboard Design. You were able to independently list key functional and non-functional requirements, sketch a high-level architecture, and think through index strategies for nearest neighbor and top-N queries.

Areas of strength included:

• Proactive in checking assumptions with the interviewer.
  
  
• Comprehensive listing of most functional and non-functional requirements.
  
  
• Ability to draft a reasonable component and data architecture.
  
  
• Thoughtful use of indexing and relational DB capabilities for ranking queries.
  
  
• Introduction of sharding for scalability as the system scales horizontally.
  
  

Opportunities for improvement:

• Clarify more assumptions up front: e.g., leaderboard dimensions (region, friends-only, time-based), score rules, deduplication behavior.
  
  
• When discussing latency, differentiate between write (ingestion) and read (rendering) paths.
  
  
• Provide peak QPS or volume estimates to guide scalability decisions.
  
  
• Be more precise with CAP tradeoffs — consider consistency guarantees at different system layers (e.g., strong consistency for top-N leaderboard).
  
  
• Refine data modeling: introduce missing entities like leaderboard, player_score, and avoid mixing unrelated fields (e.g., session_id).
  
  
• API signature corrections:
  
  
  
  • POST /score/{leaderboard_id} should return success/failure, not computed rank.
    
    
  • GET /top10/{leaderboard_id} is more appropriate than current form.
    
    
• Link your HLD design flow to functional requirements one by one for clarity.
  
  
• Explain your schema keys more clearly (e.g., upsert by leaderboard_id + user_id for scores).
  
  
• After introducing sharding, propose concrete partitioning strategies (e.g., handle celebrity leaderboard hotspots).

‍