Your question is Metrics for Imbalanced Data. 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 select metrics for imbalanced data?
Asked in the Onsite Round 3 stage. ML Machine Learning fundamentals and application round.
Focus on how you would choose primary and secondary metrics, avoid misleading accuracy, and handle thresholding, calibration, and class-specific tradeoffs. Keep your answer practical and explain how you would validate the metric choice offline and monitor it after deployment.
