Question 34
Domain 4: ML Solution Monitoring, Maintenance, and SecurityA company has deployed an ML model that detects fraudulent credit card transactions in real time in a banking application. The model uses Amazon SageMaker Asynchronous Inference. Consumers are reporting delays in receiving the inference results. An ML engineer needs to implement a solution to improve the inference performance. The solution also must provide a notification when a deviation in model quality occurs. Which solution will meet these requirements?
Correct answer: A
Explanation
Amazon SageMaker real-time inference is designed for low-latency, synchronous predictions, so it fits a banking app that needs fraud results immediately instead of waiting on "Asynchronous Inference." SageMaker Model Monitor can track model behavior and alert on drift or quality issues, providing notifications when "a deviation in model quality occurs."
Why each option is right or wrong
A. Use SageMaker real-time inference for inference. Use SageMaker Model Monitor for notifications about model quality.
Amazon SageMaker Asynchronous Inference is intended for requests that can tolerate variable latency and are queued for later processing, so it is the wrong endpoint type for a banking workflow that needs immediate fraud decisions. By contrast, SageMaker real-time inference is the synchronous endpoint option for low-latency predictions, and SageMaker Model Monitor can continuously evaluate model outputs and data quality, then publish alerts through Amazon CloudWatch when drift or quality deviation is detected.
B. Use SageMaker batch transform for inference. Use SageMaker Model Monitor for notifications about model quality.
Batch transform is for offline bulk predictions, not immediate per-transaction fraud decisions.
C. Use SageMaker Serverless Inference for inference. Use SageMaker Inference Recommender for notifications about model quality.
Inference Recommender helps choose deployment configurations, not monitor production model quality deviations.
D. Keep using SageMaker Asynchronous Inference for inference. Use SageMaker Inference Recommender for notifications about model quality.
Asynchronous inference accepts delayed responses, which conflicts with real-time fraud detection needs.