Your question is ML Model Choice and Loss. 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).
In your ML design, what model would you choose and why, and what loss function would you use?
Asked in the ML system design stage. Explain how the model and objective follow from the prediction target, available labels, serving constraints, and business costs.
Address both the initial choice and how you would validate or change it if the data, latency, or feedback characteristics differ.