Review Mode Set 1 Question 28.

[Nouman]

There is a wrong answer spotted for the following question:
A global e-commerce enterprise is building an intelligent natural language processing (NLP) system to enhance its customer engagement platform, TD-Assistance. The company uses Amazon Bedrock to access a foundation model for generative conversational responses and Amazon SageMaker AI to fine-tune a Bidirectional Encoder Representations from Transformers (BERT) model on proprietary chat transcripts for sentiment and intent analysis.

The training dataset, which contains sensitive customer messages, is stored in an Amazon S3 bucket. To improve performance, the SageMaker training job is configured for distributed training across five compute instances.

The data science team must encrypt the dataset, model checkpoints, and intermediate artifacts both at rest and in transit. Additionally, all node communications must be secure, and data transfers should never traverse the public internet.

Which combination of actions will meet these security and compliance requirements for the distributed training workload? (Select THREE.)

The corect answers are 1,4 and 6.
readonly access is not enough for model training. Sagemaker can write artifacts to s3 during the training process.

[Lois-TutorialsDojo]

Hello Nouman,

Thanks for taking the time to share your feedback on this question. We completely see your point and value your input.

We agree with your concern regarding the read-only access in Option 2. A SageMaker training job may need to read training data from Amazon S3 and also write model artifacts, checkpoints, and other outputs back to S3. Because of this, read-only permissions may not be sufficient for the training workload described.

Given this, Options 1, 4, and 6 appear to better align with the stated requirements, particularly for private S3 connectivity, communication between distributed training instances, and inter-container traffic encryption.

We’ve flagged this question for further review, and it will be updated to ensure that the options and explanation better reflect both the exam context and practical AWS best practices.

Thanks again for bringing this to our attention! This will help us improve the quality of our materials. Feel free to reach out if you have any follow-up questions.

Regards,

Lois @ Tutorials Dojo