Whenever you're ready, walk me through how you'd approach this.
How do you handle consistency versus availability in a distributed system?
Explain how you would make the choice for different operations, especially during network partitions or regional failures. Address consistency models, replication and quorum behavior, graceful degradation, recovery, and how you would validate the design in production.