AWS Certified Machine Learning Engineer Associate · Free Practice Question Medium
Question 27
A media analytics company processes raw video metadata stored in Amazon S3 buckets to generate insights on audience engagement. The company needs to create data ingestion pipelines for processing this metadata and ML model deployment pipelines to analyze and predict audience behavior patterns. The solution must efficiently handle large-scale data processing and integrate seamlessly with ML workflows.
Which solution will meet these requirements?
-
A
Use Kinesis Data Firehose to stream video metadata from Amazon S3 into the ingestion pipeline and deploy models using AWS Lambda
-
B
Use AWS Glue to preprocess raw video metadata in Amazon S3 and create a custom script to deploy models manually using SageMaker endpoints
-
C
Use AWS Glue to create data ingestion pipelines for scalable data processing and SageMaker Studio Classic to manage ML model deployment pipelines
-
D
Use SageMaker Studio Classic for both data ingestion pipelines and ML model deployment pipelines to simplify the process
Reveal correct answer
Correct answer: C
Explanation
Correct option:
Use AWS Glue to create data ingestion pipelines for scalable data processing and SageMaker Studio Classic to manage ML model deployment pipelines
AWS Glue provides serverless, scalable ETL capabilities to preprocess and transform large amounts of raw metadata from S3, enabling efficient ingestion pipelines. SageMaker Studio Classic offers an integrated environment for creating, training, deploying, and monitoring ML models. Studio Classic simplifies the process of building and managing model deployment pipelines. Therefore, the combination of AWS Glue for data ingestion and SageMaker Studio Classic for ML pipelines ensures a streamlined, scalable solution optimized for media analytics.
Incorrect options:
Use Kinesis Data Firehose to stream video metadata from Amazon S3 into the ingestion pipeline and deploy models using AWS Lambda - Kinesis Data Firehose is designed for streaming data and is not optimized for large-scale batch ingestion from Amazon S3. AWS Lambda has runtime and resource limitations, making it unsuitable for managing complex ML deployment pipelines.
Use AWS Glue to preprocess raw video metadata in Amazon S3 and create a custom script to deploy models manually using SageMaker endpoints - While AWS Glue is suitable for data preprocessing, custom scripts for model deployment require significant manual effort and are less efficient than using SageMaker Studio Classic for deployment pipelines.
Use SageMaker Studio Classic for both data ingestion pipelines and ML model deployment pipelines to simplify the process - While SageMaker Studio Classic is excellent for ML workflows, it is not designed for large-scale ETL tasks, where AWS Glue provides a better solution.
References:
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
