Your question is Production Latency and Optimization. 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 handle model optimization and inference latency when deploying large-scale models in a production environment?