Your question is Inference Under Cloud Budget. 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 would you architect a solution that balances the need for high-performance inference with strict cloud budget constraints?
Explain how you would clarify workload requirements, select model and serving strategies, and divide computation between offline and online paths. Address latency, throughput, autoscaling, hardware selection, batching, caching, model compression, availability, and cost measurement. Describe how you would evaluate quality and performance, detect feature drift and training-serving skew, and respond when the budget or latency target is exceeded.