Dali demonstrated strong technical knowledge and problem-solving abilities during our session. The approach to the algorithm question was methodical, and communicated his thought process clearly.

Areas of Strength

• Clearly enumerated core functional and non-functional requirements.
• Provided concrete NFR metrics, including thoughtful scaling estimations.
• Addressed delete-message behavior across both online and offline scenarios.
• Explained the offline → online recovery workflow, including inbox table design.
• Delivered a detailed high-level send-message architecture, covering WS server vs. chat server roles, schema design, and data flow.
• Justified database selection with strong reasoning around scalability, write throughput, and join minimization to meet availability goals.

Areas for Improvement

• This problem has a broad solution space—clarifying and aligning early with the interviewer on the exact deep-dive direction would set clearer expectations.
• Avoid diving too deep into message broadcast architecture before reaching the designated deep-dive portion.
• Add more depth on offline user detection mechanisms: heartbeats, WebSocket ping/pong, presence service with TTL, and push-delivery receipts.
• Strengthen familiarity with common messaging-system topics, such as:
  
  
  
  • Online vs. offline user handling
  • Inbox and unread-message accounting
  • Multi-tenant and isolation models
  • Multi-device sync semantics
  • Rich-media handling
  • Message search
  • Message ordering guarantees
  • Threading / conversation hierarchy

‍