Whenever you're ready, walk me through how you'd approach this.
How would you architect a system to handle high-concurrency requests while maintaining low latency?