Your question is System Design for Performance. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How do you approach designing a system with specific performance requirements?
Explain how you would clarify ambiguous requirements, translate them into measurable service-level objectives, and design an end-to-end ML system that satisfies them. Cover scale estimation, retrieval and ranking where applicable, online versus batch processing, evaluation, monitoring, and failure recovery. State which assumptions you would validate before committing to an architecture.