Your question is Scaling a System. 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).
What is scaling?
Explain how you would scale an end-to-end machine learning system as users, data volume, model complexity, and traffic increase. Cover horizontal and vertical scaling, offline training, online inference, feature serving, caching, partitioning, asynchronous processing, and graceful degradation. Discuss how you would measure capacity, latency, throughput, cost, availability, and model quality while preserving training-serving consistency.
