AIF-C01 Question 48
Single answerYou have trained a machine learning model for predicting customer churn, and now you want to deploy it to production. Your team prefers a managed solution to minimize infrastructure management efforts and ensure scalability. Which method would be the most suitable for deploying your model?
- A
Deploy the model using Amazon SageMaker Endpoints
- B
Host the model on an Amazon EC2 instance and expose it via a REST API
- C
Serve the model using an on-premises server
- D
Use Amazon S3 to directly host the model and serve predictions
Show answer and explanation
Correct answer: A
Explanation
Amazon SageMaker Endpoints are the most suitable option for deploying a machine learning model in a managed, scalable environment. SageMaker abstracts the infrastructure management and enables seamless integration with AWS services, aligning with the team's preference for a managed solution. Other options, such as EC2 or on-premises, require significant management overhead, while Amazon S3 is not a deployment service.
- A. Correct.
Amazon SageMaker Endpoints provide a fully managed service for deploying machine learning models. This option minimizes infrastructure management while providing scalability and integration with other AWS services.
- B. Incorrect.
Hosting the model on an Amazon EC2 instance requires managing the infrastructure, such as scaling and uptime, which contradicts the team's preference for a managed solution.
- C. Incorrect.
Using an on-premises server introduces significant management overhead, including hardware maintenance and scaling, which is against the team's goal of minimizing infrastructure management.
- D. Incorrect.
Amazon S3 is a storage service and cannot directly host or serve predictions for a machine learning model. It is not designed for deploying models in production.