An ML engineer wants an Amazon SageMaker notebook to automatically stop running after 1 hour of idle time.
How can the ML engineer accomplish this goal?
A. Create a lifecycle configuration in SageMaker. Copy the auto-stop-idle script from GitHub to the Start Notebook section.
B. Create a lifecycle configuration in SageMaker. Copy the auto-stop-idle script from GitHub to the Create Notebook section.
C. Track the notebook's CPU metric by using Amazon CloudWatch Logs. Invoke an AWS Lambda function from CloudWatch Logs to shut down the notebook instance if CPU utilization becomes zero.
D. Track the notebook's memory metric by using Amazon CloudWatch Logs. Invoke an AWS Lambda function from CloudWatch Logs to shut down the notebook instance if memory utilization becomes zero.
Show Answer
Correct Answer: A
Explanation: Amazon SageMaker supports automatic shutdown of idle notebook instances by using a notebook lifecycle configuration. The commonly used auto-stop-idle script is added to the Start Notebook lifecycle configuration so it runs each time the notebook instance starts and monitors idle time, stopping the instance after the configured timeout (such as 1 hour). The Create Notebook lifecycle configuration runs only once when the notebook is created, making it unsuitable for ongoing idle monitoring. CPU or memory CloudWatch-based approaches are not the recommended mechanism and do not reliably indicate notebook idleness.
Question 113
A company stores training data as a .csv file in an Amazon S3 bucket. The company must encrypt the data and must control which applications have access to the encryption key.
Which solution will meet these requirements?
A. Create a new SSH access key. Use the AWS Encryption CLI with a reference to the new access key to encrypt the file.
B. Create a new API key by using the Amazon API Gateway CreateApiKey API operation. Use the AWS CLI with a reference to the new API key to encrypt the file.
C. Create a new IAM role. Attach a policy that allows the AWS Key Management Service (AWS KMS) GenerateDataKey action. Use the role to encrypt the file.
D. Create a new AWS Key Management Service (AWS KMS) key. Use the AWS Encryption CLI with a reference to the new KMS key to encrypt the file.
Show Answer
Correct Answer: D
Explanation: AWS KMS is the AWS service designed for managing encryption keys and controlling access through key policies and IAM. Creating a customer managed KMS key and using the AWS Encryption CLI with that KMS key encrypts the data while allowing fine-grained control over which applications can use the key. The other options misuse SSH keys, API Gateway API keys, or IAM roles, which are not encryption keys.
Question 114
A company wants to build a real-time analytics application that uses streaming data from social media. An ML engineer must implement a solution that ingests and transforms 5 GB of data each minute. The solution also must load the data into a data store that supports fast queries for the real-time analytics.
Which solution will meet these requirements?
A. Use Amazon EventBridge to ingest the social media data. Use AWS Glue to transform the data. Store the transformed data in Amazon ElastiCache (Memcached).
B. Use Amazon Simple Queue Service (Amazon SQS) to ingest the social media data. Use AWS Lambda to transform the data. Store the transformed data in Amazon S3.
C. Use Amazon Simple Notification Service (Amazon SNS) to ingest the social media data. Use Amazon EMR to transform the data. Store the transformed data in Amazon RDS.
D. Use Amazon Kinesis Data Streams to ingest the social media data. Use Amazon Managed Service for Apache Flink to transform the data. Store the transformed data in Amazon DynamoDB.
Show Answer
Correct Answer: D
Explanation: Amazon Kinesis Data Streams is designed for high-throughput real-time streaming ingestion. Amazon Managed Service for Apache Flink performs stateful, low-latency stream processing and transformations. Amazon DynamoDB provides fast, low-latency queries suitable for real-time analytics workloads. The other options use services that are not designed for this scale or real-time streaming pipeline.
Question 115
A company needs to perform feature engineering, aggregation, and data preparation. After the features are produced, the company must implement a solution on AWS to process and store the features.
Which solution will meet these requirements?
A. Use Amazon SageMaker Feature Processing to process and ingest the data. Use SageMaker Feature Store to manage and store the features.
B. Use Amazon SageMaker Model Monitor to automatically ingest and transform the data. Create an Amazon S3 bucket to store the features in JSON format.
C. Use Amazon Managed Service for Apache Flink to transform the data and to ingest the data directly into Amazon SageMaker Feature Store. Use Feature Store to manage and store the features.
D. Use an Amazon SageMaker batch transform job to analyze, transform, and ingest the data. Create an Amazon DynamoDB table to store the features.
Show Answer
Correct Answer: A
Explanation: Amazon SageMaker Feature Processing is the purpose-built capability for feature engineering, aggregation, data preparation, and ingestion into SageMaker Feature Store. SageMaker Feature Store is designed to manage and store ML features for both offline training and online inference. The other options misuse services: Model Monitor is for monitoring models, Batch Transform is for inference, and Apache Flink can transform streams but is not the purpose-built end-to-end feature processing solution described.
Question 116
A company has an ML model that is deployed to an Amazon SageMaker endpoint for real-time inference. The company needs to deploy a new model. The company must compare the new model’s performance to the currently deployed model’s performance before shifting all traffic to the new model.
Which solution will meet these requirements with the LEAST operational effort?
A. Deploy the new model to a separate endpoint. Manually split traffic between the two endpoints.
B. Deploy the new model to a separate endpoint. Use Amazon CloudFront to distribute traffic between the two endpoints.
C. Deploy the new model as a shadow variant on the same endpoint as the current model. Route a portion of live traffic to the shadow model for evaluation.
D. Use AWS Lambda functions with custom logic to route traffic between the current model and the new model.
Show Answer
Correct Answer: C
Explanation: Amazon SageMaker shadow variants allow you to deploy a new model alongside the production model on the same endpoint. The production model serves responses to users while the shadow model receives a copy of live requests for performance evaluation without affecting users. This is the built-in, lowest-operational-effort approach for comparing models before shifting traffic.
Question 117
A company is exploring generative AI and wants to add a new product feature. An ML engineer is making API calls from existing Amazon EC2 instances to Amazon Bedrock. The EC2 instances are in a private subnet and must remain private during the implementation. The EC2 instances have an assigned security group that allows access to all IP addresses in the private subnet.
What should the ML engineer do to establish a connection between the EC2 instances and Amazon Bedrock?
A. Modify the security group to allow inbound and outbound traffic to and from Amazon Bedrock.
B. Use AWS PrivateLink to access Amazon Bedrock through an interface VPC endpoint.
C. Configure Amazon Bedrock to use the private subnet where the EC2 instances are deployed.
D. Link the existing VPC to Amazon Bedrock by using an AWS Direct Connect connection.
Show Answer
Correct Answer: B
Explanation: Amazon Bedrock is an AWS managed service. To allow EC2 instances in a private subnet to access Bedrock without using the public internet, create an interface VPC endpoint for Amazon Bedrock using AWS PrivateLink. Security groups alone cannot create connectivity to a managed service, Bedrock cannot be deployed into your subnet, and AWS Direct Connect is for on-premises connectivity, not VPC-to-managed-service access.
Question 118
A company wants to launch a new internal generative AI interface to answer user questions. The interface will be based on a popular open source large language model (LLM).
Which combination of steps will deploy the interface with the LEAST operational overhead? (Choose two.)
A. Use Amazon SageMaker JumpStart to deploy the LLM.
B. Download the LLM as a .zip file. Deploy the LLM on a GPU-based Amazon EC2 instance.
C. Create a frontend HTML interface that uses an Amazon API Gateway WebSocket API with AWS Lambda functions to handle the user interaction.
D. Use Amazon QuickSight to create a UI to handle the user interaction.
E. Use Amazon Lex to create a UI to handle the user interaction.
Show Answer
Correct Answer: A, E
Explanation: Amazon SageMaker JumpStart provides managed deployment of popular open source LLMs, minimizing infrastructure management compared with manually provisioning GPU EC2 instances. Amazon Lex is a fully managed conversational interface that handles user interaction, session management, and integration with backend AI services, resulting in less operational overhead than building and maintaining a custom API Gateway WebSocket and Lambda frontend. QuickSight is for business intelligence dashboards, not conversational AI interfaces.
Question 119
A company is using Amazon EMR. The company has a large dataset in Amazon S3 that needs to be ingested into Amazon SageMaker Feature Store. The dataset contains historical data and real-time streaming data.
The company must ensure that the Feature Store online store is updated with the most recent data as soon as the data becomes available. The company also must maintain a complete Feature Store offline store for batch processing.
Which solution will meet these requirements?
A. Use the PutRecord API in Feature Store Runtime to ingest all the data into the online store.
B. Use the PutRecord API in Feature Store Runtime to ingest all the data into the offline store.
C. Use the Feature Store Spark connector to ingest the data as Spark DataFrames with the online store and offline store enabled.
D. Use the Feature Store Spark connector to ingest the data as Spark DataFrames with only the online store enabled.
Show Answer
Correct Answer: C
Explanation: The Feature Store Spark connector is designed for high-throughput ingestion from Apache Spark workloads such as Amazon EMR. Enabling both the online and offline stores writes records to the online store for low-latency serving while maintaining the complete offline store for historical analytics and batch processing. PutRecord is intended for individual record ingestion rather than efficient bulk Spark-based ingestion.
Question 120
A company is developing a new online application to gather information from customers. An ML engineer has developed a new ML model that will determine a score for each customer. The model will use the score to determine which product to display to the customer. The ML engineer needs to minimize response-time latency for the model.
How should the ML engineer deploy the application in Amazon SageMaker to meet these requirements?
A. Configure batch transform.
B. Configure a real-time inference endpoint.
C. Configure a serverless inference endpoint.
D. Configure an asynchronous inference endpoint.
Show Answer
Correct Answer: B
Explanation: A customer-facing application that performs synchronous per-request predictions and requires minimal response-time latency should use a SageMaker real-time inference endpoint. Batch transform is for offline batch predictions, asynchronous inference is for long-running requests where latency is not critical, and serverless inference is better for intermittent or unpredictable traffic but generally does not provide the lowest latency compared with provisioned real-time endpoints.
Question 121
An ML engineer is deploying a trained model to an Amazon SageMaker endpoint. The ML engineer needs to receive alerts when data quality issues occur in production.
Which solution will meet this requirement?
A. Configure an Amazon CloudWatch metric alarm and a corresponding action to send an Amazon Simple Notification Service (Amazon SNS) notification.
B. Integrate the SageMaker endpoint with a SageMaker Clarify processing job. Configure an Amazon CloudWatch alarm to provide alerts.
C. Configure a monitoring job in SageMaker Model Monitor. Integrate Model Monitor with Amazon CloudWatch to provide alerts.
D. Configure a data flow in SageMaker Data Wrangler. Integrate Data Wrangler with Amazon CloudWatch to provide alerts.
Show Answer
Correct Answer: C
Explanation: Amazon SageMaker Model Monitor is designed to continuously monitor production endpoints for data quality, model quality, bias, and feature attribution drift. It publishes metrics to Amazon CloudWatch, where alarms can be configured to trigger notifications (for example, via Amazon SNS) when data quality violations occur. CloudWatch alarms alone do not detect data quality issues, Clarify focuses on bias/explainability, and Data Wrangler is for data preparation rather than production monitoring.
$19
Get all 235 questions with detailed answers and explanations
Instant download HTML + PDF delivered the moment payment clears.
Secure Stripe checkout we never see or store your card details.
7-day refund if files are defective see our refund policy.