Your question is Evaluating Alternative ML Metrics. 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 metrics would you use to evaluate alternative machine learning techniques?
Answer as if you are selecting between models for a production use case. Focus on how you would compare performance, validate improvements, and avoid being misled by a single metric.