Welcome to your interview.
The question is on your right: KV Cache for Multi-Tenant Serving. Take a moment with it first.
Talk your thinking through with me if you like - when you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes). Discussion and graded submissions share your five interviewer interactions, so spend them well.
How would you implement and optimize a KV cache in a multi-tenant inference cluster to prevent memory fragmentation?