MLA-C01 AWS Certified Machine Learning Engineer - Associate Questions and Answers

AWS Key Management Service (AWS KMS) is the recommended service for encryption and key access control. By creating a customer-managed KMS key, the company can define granular IAM policies that control which applications and roles can use the key.

The AWS Encryption CLI integrates directly with KMS and enables client-side encryption of files before storing them in Amazon S3. This approach ensures data is encrypted at rest and that only authorized principals can decrypt it.

SSH keys and API keys are not designed for data encryption. IAM roles alone do not create or manage encryption keys—they only grant permissions.

AWS documentation explicitly states that KMS customer-managed keys provide centralized key management, auditing, and access control.

Therefore, Option D is the correct and AWS-aligned solution.

Problem Description:

The F1 score, which is a balance of precision and recall, has decreased significantly. This indicates the model ' s predictions are no longer aligned with the real-world data distribution.

Why Concept Drift?

Concept drift occurs when the statistical properties of the target variable or features change over time. For example, customer behaviors or subscription cancellation patterns may have shifted, leading to reduced model accuracy.

Signs of Concept Drift:

Deviation in performance metrics (e.g., F1 score) over time.

Declining prediction accuracy for certain groups or scenarios.

Solution:

Monitor for drift using tools like SageMaker Model Monitor.

Regularly retrain the model with updated data to account for the drift.

Why Not Other Options?:

B: Model complexity is unrelated if the model initially performed well.

C: Data quality issues would have been detected during baseline analysis.

D: Incorrect ground truth labels would have resulted in a consistently poor baseline.

Conclusion: The decrease in F1 score is most likely due to concept drift in the customer data, requiring retraining of the model with new data.

Step 1: An S3 event notification invokes the pipeline when new data is uploaded.

Step 2: SageMaker retrains the model by using the data in the S3 bucket.

Step 3: The pipeline deploys the model to a SageMaker endpoint.

Step 1: An S3 Event Notification Invokes the Pipeline When New Data is Uploaded

Why? The CI/CD pipeline should be triggered automatically whenever new training data is uploaded to Amazon S3. S3 event notifications can be configured to send events to AWS services like Lambda, which can then invoke AWS CodePipeline.

How? Configure the S3 bucket to send event notifications (e.g., s3:ObjectCreated:*) to AWS Lambda, which in turn triggers the CodePipeline.

Step 2: SageMaker Retrains the Model by Using the Data in the S3 Bucket

Why? The uploaded data is used to retrain the ML model to incorporate new information and maintain performance. This step is critical to updating the model with fresh data.

How? Define a SageMaker training step in the CI/CD pipeline, which reads the training data from the S3 bucket and retrains the model.

Step 3: The Pipeline Deploys the Model to a SageMaker Endpoint

Why? Once retrained, the updated model must be deployed to a SageMaker endpoint to make it available for real-time inference.

How? Add a deployment step in the CI/CD pipeline, which automates the creation or update of the SageMaker endpoint with the retrained model.

Order Summary:

An S3 event notification invokes the pipeline when new data is uploaded.

SageMaker retrains the model by using the data in the S3 bucket.

The pipeline deploys the model to a SageMaker endpoint.

This configuration ensures an automated, efficient, and scalable CI/CD pipeline for continuous retraining and deployment of the ML model in Amazon SageMaker.

AWS documentation strongly recommends using custom Docker containers when ML workloads require consistent access to custom dependencies across processing jobs, training jobs, and pipelines.

By building a single Docker image that contains all required libraries and hosting it in Amazon ECR, the ML engineer ensures that every SageMaker job uses the same runtime environment. This approach eliminates the need for repetitive installation steps and avoids environment drift.

Manually installing libraries in managed containers is error-prone and not reusable across jobs. Notebook instances are not designed to host production jobs and pipelines. Running code externally breaks the SageMaker workflow and increases operational complexity.

Using a custom container is a one-time setup that provides maximum reuse with minimal ongoing effort, making it the least implementation effort option in the long run.

Therefore, Option B is the correct and AWS-recommended answer.

The key requirements are real-time processing, high throughput, and minimal operational overhead. Amazon Kinesis Data Streams is designed for ingesting thousands of events per second with low latency.

For anomaly detection on streaming data, Amazon Managed Service for Apache Flink provides a built-in Random Cut Forest (RCF) function. RCF is an unsupervised anomaly detection algorithm that works well on numerical streaming data and does not require labeled training data.

This fully managed combination eliminates the need to deploy or maintain SageMaker endpoints, EC2 instances, or custom ML pipelines. Options B and C introduce unnecessary infrastructure and model management overhead. Option D is batch-oriented and unsuitable for real-time anomaly detection.

Therefore, using Kinesis Data Streams with Flink’s built-in Random Cut Forest is the most scalable and low-overhead solution.

Amazon SageMaker AI automatically collects aggregated metadata from training jobs to improve service reliability, performance, and operational insights. This metadata can include information such as algorithm usage, instance types, resource utilization, and job configuration details. However, AWS documentation clearly states that customers can opt out of SageMaker metadata collection to meet regulatory or compliance requirements.

SageMaker provides a supported mechanism to disable metadata tracking at the training job level. By explicitly opting out of metadata tracking when submitting training jobs—either through the AWS Management Console, AWS CLI, or SDK—the service will stop collecting aggregated metadata for those jobs. This option is specifically designed for customers with strict compliance, data residency, or regulatory constraints.

Option B is incorrect because running training jobs in a private subnet within a custom VPC controls network isolation, not service-level telemetry or metadata collection. Metadata collection occurs at the SageMaker service layer and is independent of VPC configuration.

Option C is also incorrect because encrypting training data with a customer-managed AWS KMS key protects data at rest and in transit but does not prevent SageMaker from collecting operational metadata about training jobs.

Option D is incorrect because AWS Nitro instances provide enhanced security and performance isolation at the infrastructure level but have no impact on SageMaker’s metadata collection mechanisms.

Therefore, opting out of metadata tracking for training jobs is the only solution that directly addresses the compliance requirement and is explicitly supported by AWS documentation.

In Amazon SageMaker, distributed training and distributed processing jobs often involve multiple instances exchanging data over the network. By default, when these jobs run inside a VPC, network traffic remains private but is not automatically encrypted between instances. When compliance or security requirements mandate encryption of in-transit data, additional configuration is required.

The correct solution is to enable inter-container traffic encryption, which ensures that all network communication between containers running on different instances is encrypted using TLS. Amazon SageMaker provides a built-in feature for this purpose. When inter-container traffic encryption is enabled, SageMaker automatically configures secure communication channels between all nodes participating in a distributed job, including training clusters and processing jobs.

Option A (Network isolation) is incorrect because network isolation prevents containers from making outbound network calls and accessing the internet. It does not encrypt traffic between instances.

Option B (Security groups) is incorrect because security groups control network access and traffic flow, not encryption. They can restrict which instances can communicate, but they do not provide data-in-transit encryption.

Option D (VPC flow logs) is incorrect because VPC flow logs are used for monitoring and auditing network traffic, not for encrypting it.

AWS documentation explicitly states that enabling inter-container traffic encryption is the recommended and supported approach for encrypting data exchanged between instances during distributed SageMaker jobs. This feature aligns with enterprise security best practices and regulatory requirements for protecting sensitive ML training data in transit.

Therefore, Option C is the only solution that directly fulfills the encryption requirement for distributed SageMaker workloads.

Step 1

Create an IAM role for Amazon Redshift to use to access the Data Catalog and the S3 bucket that contains underlying data.

This role must include:

Permissions for AWS Glue Data Catalog (e.g., glue:GetDatabase, glue:GetTables)

Permissions for the Amazon S3 bucket that stores the underlying data

Step 2

Attach the IAM role to the Redshift namespace.

Redshift Serverless uses a namespace, not a cluster, so the role must be associated with the namespace to allow Redshift to assume it when querying external data.

Step 3

Create an external schema in Amazon Redshift to point to the Data Catalog database.

The external schema maps Redshift to the Glue Data Catalog database so Redshift can query the tables stored in S3.

Amazon SageMaker automatically tracks the Invocations metric, which represents the number of API calls made to the endpoint, in Amazon CloudWatch. By adding this metric to a CloudWatch dashboard, you can monitor the endpoint ' s activity in real-time. Setting up a CloudWatch alarm allows the system to send notifications whenever the API call events exceed the defined threshold, meeting both the monitoring and notification requirements efficiently.

For Amazon EMR, the primary node and core nodes handle the critical functions of the cluster, including data storage (HDFS) and processing. Running them on On-Demand Instances ensures high availability and prevents data loss, as Spot Instances can be interrupted. The task nodes, which handle additional processing but do not store data, can use Spot Instances to reduce costs without compromising the cluster ' s resilience or data integrity. This configuration balances cost-effectiveness and reliability.

 Why Linear Learner?

SageMaker ' s Linear Learner algorithm is well-suited for binary classification problems such as fraud detection. It handles class imbalance effectively by incorporating built-in options for weight balancing across classes.

Linear Learner can capture patterns in the data while being computationally efficient.

 Key Features of Linear Learner:

Automatically weights minority and majority classes.

Supports both classification and regression tasks.

Handles interdependencies among features effectively through gradient optimization.

 Steps to Implement:

Use the SageMaker Python SDK to set up a training job with the Linear Learner algorithm.

Configure the hyperparameters to enable balanced class weights.

Train the model with the balanced dataset created using SageMaker Data Wrangler.

Linear regression model whose coefficients should shrink but not become zero

Answer: L2 regularization

AWS says L2 produces smaller overall weight values and is the right fit when coefficients should be reduced without being forced to zero.

Polynomial regression model with irrelevant polynomial terms that should be eliminated

Answer: L1 regularization

AWS says L1 reduces the number of features used by pushing small weights to zero, which matches elimination of irrelevant terms.

Logistic regression model that has highly correlated features to eliminate highly redundant predictors

Answer: L1 regularization

This is the nuanced one. AWS says L2 stabilizes weights when there is high correlation between features, but because the question explicitly says eliminate highly redundant predictors, L1 is the better match since it creates sparsity and removes predictors by zeroing coefficients.

When running consecutive training jobs in Amazon SageMaker, infrastructure provisioning can introduce latency, as each job typically requires the allocation and setup of compute resources. To minimize this startup time and enhance efficiency, Amazon SageMaker offers Managed Warm Pools.

Key Features of Managed Warm Pools:

Reduced Latency: Reusing existing infrastructure significantly reduces startup time for training jobs.

Configurable Retention Period: Allows retention of resources after training jobs complete, defined by the KeepAlivePeriodInSeconds parameter.

Automatic Matching: Subsequent jobs with matching configurations (e.g., instance type) can reuse retained infrastructure.

Implementation Steps:

Request Warm Pool Quota Increase: Increase the default resource quota for warm pools through AWS Service Quotas.

Configure Training Jobs:

Set KeepAlivePeriodInSeconds for the first training job to retain resources.

Ensure subsequent jobs match the retained pool ' s configuration to enable reuse.

Monitor Warm Pool Usage: Track warm pool status through the SageMaker console or API to confirm resource reuse.

Considerations:

Billing: Resources in warm pools are billable during the retention period.

Matching Requirements: Jobs must have consistent configurations to use warm pools effectively.

Alternative Options:

Managed Spot Training: Reduces costs by using spare capacity but doesn’t address startup latency.

SageMaker Training Compiler: Optimizes training time but not infrastructure setup.

SageMaker Distributed Data Parallelism Library: Enhances training efficiency but doesn’t reduce setup time.

By using Managed Warm Pools, the company can significantly reduce startup latency for consecutive training jobs, ensuring faster experimentation cycles with minimal operational overhead.

AWS Documentation: Managed Warm Pools

AWS Blog: Reduce ML Model Training Job Startup Time

Problem Description:

The dataset includes multiple data sources:

Transaction logs and customer profiles in Amazon S3.

Tables in an on-premises MySQL database.

There is a class imbalance in the dataset and interdependencies among features that need to be addressed.

The solution requires data aggregation from diverse sources for centralized processing.

Why AWS Lake Formation?

AWS Lake Formation is designed to simplify the process of aggregating, cataloging, and securing data from various sources, including S3, relational databases, and other on-premises systems.

It integrates with AWS Glue for data ingestion and ETL (Extract, Transform, Load) workflows, making it a robust choice for aggregating data from Amazon S3 and on-premises MySQL databases.

How It Solves the Problem:

Data Aggregation: Lake Formation collects data from diverse sources, such as S3 and MySQL, and consolidates it into a centralized data lake.

Cataloging and Discovery: Automatically crawls and catalogs the data into a searchable catalog, which the ML engineer can query for analysis or modeling.

Data Transformation: Prepares data using Glue jobs to handle preprocessing tasks such as addressing class imbalance (e.g., oversampling, undersampling) and handling interdependencies among features.

Security and Governance: Offers fine-grained access control, ensuring secure and compliant data management.

Steps to Implement Using AWS Lake Formation:

Step 1: Set up Lake Formation and register data sources, including the S3 bucket and on-premises MySQL database.

Step 2: Use AWS Glue to create ETL jobs to transform and prepare data for the ML pipeline.

Step 3: Query and access the consolidated data lake using services such as Athena or SageMaker for further ML processing.

Why Not Other Options?

Amazon EMR Spark jobs: While EMR can process large-scale data, it is better suited for complex big data analytics tasks and does not inherently support data aggregation across sources like Lake Formation.

Amazon Kinesis Data Streams: Kinesis is designed for real-time streaming data, not batch data aggregation across diverse sources.

Amazon DynamoDB: DynamoDB is a NoSQL database and is not suitable for aggregating data from multiple sources like S3 and MySQL.

Conclusion: AWS Lake Formation is the most suitable service for aggregating data from S3 and on-premises MySQL databases, preparing the data for downstream ML tasks, and addressing challenges like class imbalance and feature interdependencies.

AWS Lake Formation Documentation

AWS Glue for Data Preparation

AWS provides Amazon SageMaker Data Wrangler as a native tool for importing, transforming, and analyzing data from multiple sources directly into SageMaker Studio. Data Wrangler supports Amazon S3, Amazon Athena, and Snowflake as built-in data sources through managed connectors.

Using Data Wrangler, ML engineers can query data from Athena using SQL, load structured files from S3, and securely connect to Snowflake without writing custom ingestion code. This approach significantly reduces development effort and aligns with AWS best practices for rapid ML experimentation.

Option A is incorrect because AWS Glue DataBrew is designed for data preparation but does not natively integrate with SageMaker training workflows. Option B introduces unnecessary complexity and is not intended for direct ML data loading. Option C focuses on feature storage, not raw historical data ingestion.

Therefore, SageMaker Data Wrangler is the correct solution.

The correct answer is C. Download the training data to the estimator by using fast file mode. Point the data loader to the location specified by the S3 path.

When training neural network models in Amazon SageMaker, I/O operations can become a bottleneck, especially when reading large datasets directly from S3. SageMaker provides a fast file mode that allows the training job to cache data locally on the compute instance before training. This significantly reduces data-loading latency, as the model can access the data from local storage instead of repeatedly fetching it over the network from S3.

Fast file mode automatically handles staged caching of S3 data in the container’s file system or attached EBS volumes, ensuring that the estimator sees high-throughput access to the dataset. This approach improves overall training speed without requiring manual configuration of EBS or EFS volumes.

Option A, provisioning EBS with io1 storage, may increase I/O performance but requires manual copying of the data and adds operational complexity. Option B, using EFS, introduces network-based I/O overhead and is not optimized for high-throughput sequential reads, which are typical in neural network training. Option D, changing the S3 path to a hyperparameter, has no effect on I/O performance or data-loading speed—it is merely a way to pass metadata to the training job.

Using fast file mode is AWS-recommended when profiling shows that training is bottlenecked by data access. It provides a managed, high-throughput, low-latency caching layer, allowing ML engineers to focus on model architecture and hyperparameter tuning rather than data transfer optimizations. This aligns with best practices for ML model development by reducing training time while maintaining scalability and simplicity in SageMaker training workflows.

For near real-time anomaly detection on streaming IoT data, AWS recommends Amazon Managed Service for Apache Flink. Flink is designed for stateful, low-latency stream processing and is well suited for time-series sensor analytics.

A key requirement is application resilience during deployments. Managed Flink supports system rollback, which automatically reverts to the last stable application version if a new deployment fails. This capability ensures uninterrupted processing even if application bugs are introduced, meeting the company’s reliability requirement.

Manual rollback (Option B) introduces operational risk and delays. Amazon Data Firehose (Options C and D) is a delivery service and does not support complex anomaly detection logic or application version rollback.

Therefore, using Managed Flink with system rollback enabled is the correct solution.

SageMaker JumpStart provides access to pre-trained models, including large language models (LLMs), which can be easily deployed and fine-tuned with a low-code/no-code (LCNC) approach. Using SageMaker Autopilot with JumpStart simplifies the fine-tuning process by automating model optimization and reducing the need for extensive coding, making it the ideal solution for this requirement.

The correct answer is A. The AWS Glue service quotas have been reached. AWS Glue is commonly used in ML data preparation pipelines to perform ETL operations before model training or feature engineering. When AWS Glue workloads exceed account-level or Region-level service quotas, jobs can be delayed, queued, throttled, or fail. AWS documentation states that AWS Glue has service quotas, also called limits, for resources and operations in each account and Region. These include quotas for concurrent compute capacity measured in DPUs, crawlers, jobs, triggers, queued job runs, and other Glue resources.

The key clue in the question is “throttling errors in the ETL jobs.” AWS specifically explains that Glue API requests are throttled on a per-account, per-Region basis to maintain service performance. If the team is running many jobs, processing a large amount of sensor data, or starting too many Glue job runs concurrently, the workload can exceed the permitted quota and cause throttling.

Option B is less likely because insufficient network bandwidth between IoT devices and the AWS Region would mainly explain slower ingestion into S3, but it would not directly explain AWS Glue ETL throttling errors. Option C might cause poor job performance, but lack of parallel optimization usually causes long runtimes, not service-level throttling. Option D is incorrect because missing Amazon S3 permissions would typically cause access-denied or authorization failures, not throttling. Therefore, the best cause is that the AWS Glue service quotas, such as concurrent job runs, API rate limits, or DPU capacity limits, have been reached.

The company already has custom Python training scripts, proprietary datasets, and uses PyTorch, with significant domain-specific logic embedded in the model. The goal is to migrate these workloads to AWS with the least development effort.

According to AWS documentation, Amazon SageMaker AI script mode is explicitly designed for this scenario. Script mode allows customers to bring their existing training scripts with minimal or no code changes and run them using prebuilt SageMaker framework containers, including PyTorch. This approach eliminates the need to redesign models or rewrite training logic while still benefiting from SageMaker’s managed infrastructure, scalability, monitoring, and security.

Option A is incorrect because SageMaker built-in algorithms require adapting data formats and training logic to AWS-provided implementations, which would increase development effort and may not support proprietary domain logic.

Option C is also incorrect because building and maintaining a custom container requires additional effort for containerization, dependency management, security updates, and lifecycle maintenance—making it more complex than necessary.

Option D is not viable because purchasing models from AWS Marketplace would not support the company’s proprietary datasets or unique domain knowledge embedded in existing models.

Therefore, using SageMaker AI script mode with prebuilt PyTorch containers is the fastest, most efficient, and AWS-recommended migration path that minimizes development effort while preserving existing workflows.

AWS Glue Data Quality provides managed, declarative data quality checks with minimal configuration. Combined with Glue crawlers, it enables automatic schema discovery and quality validation without custom code.

Option A uses native AWS services designed for this exact purpose, minimizing operational overhead. Options B and C require custom code and maintenance. Option D is not designed for data validation.

AWS documentation explicitly recommends Glue Data Quality rules for scalable, automated data quality checks in analytics pipelines.

Therefore, Option A is the correct and AWS-aligned solution.

The requirement that each parameter trial is built based on the performance of the previous trial is the defining characteristic of Bayesian optimization. In Amazon SageMaker Automatic Model Tuning, Bayesian optimization uses prior trial results to intelligently select the next set of hyperparameters, making it far more efficient than grid or random search—especially in large, mixed search spaces.

This scenario includes both categorical (20) and continuous (7) hyperparameters and allows up to 1,000 training jobs, which is well within the supported limits of SageMaker AMT. Bayesian optimization natively supports mixed parameter types and is explicitly recommended by AWS for large, high-dimensional search spaces where exhaustive grid search is impractical.

Option A and D (grid search) do not meet the requirement because grid search evaluates combinations independently and does not learn from previous trials. Additionally, grid search becomes computationally infeasible as dimensionality increases.

Option B (random search) also evaluates trials independently and does not leverage previous results, violating the core requirement.

Therefore, defining both categorical and continuous parameters and using Bayesian optimization with a maximum of 1,000 jobs is the correct and AWS-recommended solution.

Apache Parquet is a columnar storage file format optimized for complex and large datasets. It provides efficient reading and processing by accessing only the required columns, which reduces I/O and speeds up data handling. This makes it ideal for use with Amazon SageMaker Canvas, where minimizing processing time is important for training ML models. Parquet is also compatible with S3 and widely supported in data analytics and ML workflows.

Option A is correct because Amazon Rekognition provides a built-in EyeDirection attribute that indicates the direction a person’s eyes are gazing, expressed through pitch and yaw. AWS documentation describes EyeDirection as showing where the eyes are looking independently of head pose. For a driver-monitoring use case, this is directly useful for identifying whether a driver is looking away from the road and may be distracted.

AWS also announced that Rekognition’s eye gaze direction detection is intended to support safety use cases and to help identify where users focus their attention. That makes it a strong match for monitoring in-vehicle cameras for distraction detection. Because Rekognition is a fully managed service with ready-to-use computer vision capabilities, it requires far less operational effort than collecting data, labeling it, training a custom model in SageMaker, and maintaining that model over time.

The other options are less appropriate. Option B could work technically, but building and maintaining a custom SageMaker model would require significantly more engineering and operational work than using a managed Rekognition feature. Option C introduces unnecessary complexity by adding a third-party system when AWS already provides a directly relevant managed capability. Option D is unrelated because Amazon Comprehend analyzes text, while the input here is camera footage. Since the question asks for the solution with the least operational effort, the best AWS-aligned answer is to use the managed Rekognition eye gaze detection capability. Therefore, the fully verified answer is A.

The SageMaker Canvas user needs permissions to access the Amazon S3 bucket where the model artifacts are stored to retrieve the model for use in Canvas.

Registering the model in the SageMaker Model Registry allows the model to be tracked and managed within the SageMaker ecosystem. This makes it accessible for tuning and deployment through SageMaker Canvas.

This combination ensures proper access control and integration within SageMaker, enabling the Canvas user to work with the model.

Amazon Data Firehose supports near real-time delivery by configuring the buffer interval and buffer size. AWS documentation states that setting the buffer interval to the minimum (as low as 1 second) enables low-latency ingestion.

By using zero or minimal buffering and tuning PutRecordBatch, data is delivered to OpenSearch almost immediately, enabling sub-second dashboard updates.

DataSync and SQS are not designed for real-time streaming analytics. Increasing the buffer interval worsens latency.

AWS explicitly recommends Firehose with minimal buffering for real-time OpenSearch ingestion.

Therefore, Option A is the correct and AWS-verified solution.

This issue occurs because target tracking auto scaling allows the endpoint to scale down to zero, and scaling up only happens after traffic arrives. At the start of the business day, no instances are running, so the first requests experience cold-start delays.

AWS documentation for Amazon SageMaker recommends using scheduled or step scaling policies when predictable traffic patterns exist. In this case, business hours are predictable, so the best practice is to proactively scale the endpoint before traffic arrives.

Option D correctly uses Amazon CloudWatch alarms with step scaling to increase the minimum instance count from zero to one at the start of the business day. This ensures at least one warm instance is ready to handle requests immediately, eliminating startup latency.

Options A, B, and C do not guarantee instance availability before traffic begins. Cooldown tuning and metric changes only react after load is detected.

Therefore, scheduled step scaling using CloudWatch alarms is the correct solution.

To minimize communication overhead during distributed training:

1. Same VPC Subnet: Ensures low-latency communication between training instances by keeping the network traffic within a single subnet.

2. Same AWS Region and Availability Zone: Reduces network latency further because cross-AZ communication incurs additional latency and costs.

3. Data in the Same Region and AZ: Ensures that the training data is accessed with minimal latency, improving performance during training.

This configuration optimizes communication efficiency and minimizes overhead.

In Amazon QuickSight anomaly detection, the Direction parameter controls which deviations from the expected baseline are flagged as anomalies. When Direction is set to All, QuickSight detects both higher-than-expected and lower-than-expected values. This results in a larger set of detected anomalies, increasing recall.

When the Direction parameter is changed to Lower than expected, the anomaly detection algorithm filters out all higher-than-expected anomalies and only reports unusually low values. Because the scope of detection is narrowed, the total number of detected anomalies decreases. As a result, fewer true anomalies are identified overall, which reduces recall.

Since fewer data points are evaluated as potential anomalies, the frequency of anomaly identification also decreases. This behavior is consistent with AWS documentation, which explains that directional filtering limits detection to a subset of deviations and therefore reduces overall anomaly counts.

Therefore, changing the Direction from All to Lower than expected leads to decreased anomaly identification frequency and decreased recall.

Dynamic data masking allows you to control how sensitive data is presented to users at query time, without modifying or storing transformed versions of the source data. Amazon Redshift supports dynamic data masking, which can be implemented with minimal effort. This solution ensures that the data scientist can access the required information while sensitive data remains protected, meeting the requirements efficiently and with the least implementation effort.

Option A is correct because AWS documentation states that regularization helps prevent linear models from overfitting, and specifically that L1 regularization reduces the number of features used in the model by pushing small feature weights to zero. AWS further explains that L1 regularization produces sparse models and reduces noise. That is the exact combination needed in this scenario: the model is overfitting, and the data contains unnecessary features whose impact should be reduced.

AWS guidance on overfitting also says that to reduce model overfitting, you should reduce model flexibility by using feature selection and increasing the amount of regularization. L1 regularization is especially suitable here because it does both in practice: it acts as a regularizer and also effectively performs feature selection by shrinking unhelpful feature coefficients to zero. Even though the option wording says “apply L1 regularization to the training data,” the only answer choice that aligns with AWS-documented techniques for both overfitting reduction and unnecessary-feature suppression is L1 regularization with retraining.

The other options do not meet both requirements. SageMaker Debugger monitors and helps analyze training jobs, but it is not the feature that “applies L1 regularization” to a running model. Increasing training iterations can make overfitting worse, not better. Decreasing training iterations may reduce overfitting in some cases, but it does not specifically address the presence of unnecessary features. Therefore, the AWS-aligned choice that best reduces overfitting and the impact of irrelevant features is A.

Amazon SageMaker Pipelines is designed for creating, automating, and managing end-to-end ML workflows, including complex data preprocessing tasks. It supports handling large datasets and can integrate with custom steps, such as NLP transformations. By combining SageMaker Pipelines with Amazon EventBridge, the entire workflow can be triggered and automated efficiently, meeting the requirements for scalability, automation, and processing complexity.

AWS recommends shadow testing to evaluate a new model against a production model with minimal operational overhead. Using production variants on a single SageMaker endpoint allows traffic to be routed to multiple models without managing additional endpoints.

With a shadow variant, the new model receives a copy of live traffic but does not affect production responses. Performance metrics such as latency, accuracy, and error rates can be compared directly against the current model using Amazon CloudWatch metrics. This approach is natively supported by Amazon SageMaker Endpoints.

Options A, B, and D introduce unnecessary complexity by requiring additional endpoints, traffic routing infrastructure, or custom code.

Therefore, deploying the new model as a shadow variant on the same endpoint is the most efficient solution.

Amazon MSK Serverless provides a fully managed Apache Kafka-compatible service that automatically handles provisioning, scaling, and capacity management. AWS documentation states that MSK Serverless is designed for customers who want Kafka functionality without managing infrastructure.

Option B requires capacity planning and scaling management. Option C uses proprietary technology rather than open source. Option D requires full infrastructure management.

MSK Serverless delivers near real-time streaming with minimal operational overhead while maintaining compatibility with open-source Kafka tooling.

Therefore, Option A is the correct solution.

The requirement is event-driven automation when new data arrives in Amazon S3, followed by training and model registration. Amazon EventBridge natively supports S3 object creation events and can trigger downstream workflows immediately.

By using EventBridge to start an AWS Step Functions workflow that includes a training step and a SageMaker Model Registry registration step, the pipeline runs automatically as soon as new data is uploaded—well within the 24-hour requirement.

Option A introduces unnecessary polling and delay. Option B is time-based and does not ensure alignment with data readiness. Option C is invalid because S3 Lifecycle rules manage object transitions, not workflow execution.

Therefore, EventBridge-triggered Step Functions is the correct solution.

AWS Glue DataBrew is designed specifically for no-code and low-code data preparation, making it the fastest and lowest-overhead solution for resolving common data quality issues. DataBrew provides built-in transformations for deduplication, missing value imputation, and outlier handling while preserving data integrity.

Option A uses standard statistical techniques such as median imputation and value normalization, which are widely accepted and maintain the distribution of the data. DataBrew jobs are fully managed and do not require infrastructure setup or maintenance.

Option B deletes records, which can lead to data loss and does not preserve integrity. Option C introduces unnecessary infrastructure complexity and uses poor data imputation practices. Option D provides advanced capabilities but requires more configuration and ML expertise, increasing operational overhead.

AWS documentation clearly positions DataBrew as the preferred solution for quick, reliable data cleaning with minimal effort.

Therefore, Option A is the correct answer.

AWS provides Amazon SageMaker Model Monitor as a native solution for detecting data drift and model quality issues in production ML pipelines. Model Monitor continuously analyzes incoming inference data and compares it with baseline training data to identify schema drift, feature distribution drift, and data quality anomalies.

When drift thresholds are violated, Model Monitor generates CloudWatch metrics and alerts. These alerts can directly trigger an AWS Lambda function, which can then programmatically initiate a SageMaker retraining job or start a SageMaker Pipeline execution. This design is explicitly documented by AWS as the recommended architecture for automated retraining workflows.

Option A is incorrect because AWS Glue is a data integration service and does not provide ML-specific drift detection capabilities.

Option B is incorrect because Apache Flink is designed for stream processing, not ML data drift detection.

Option D is incorrect because Amazon QuickSight anomaly detection is intended for business intelligence metrics, not ML feature drift.

Therefore, using SageMaker Model Monitor with AWS Lambda automation is the correct, AWS-native solution for drift-driven retraining.

For real-time anomaly detection on streaming data, AWS recommends using Amazon Managed Service for Apache Flink, which provides serverless, scalable stream processing with built-in ML functions.

Apache Flink includes a native RANDOM_CUT_FOREST (RCF) function for unsupervised anomaly detection. This eliminates the need to deploy, manage, or scale custom ML endpoints. When combined with Amazon Kinesis Data Streams, the solution can process thousands of records per second with very low latency.

Option B introduces multiple services and custom logic, increasing operational overhead. Option C requires managing EC2 infrastructure and Kafka clusters. Option D is batch-oriented and does not meet real-time requirements.

AWS documentation explicitly highlights Kinesis + Managed Flink + RCF as the lowest-overhead solution for real-time anomaly detection.

Therefore, Option A is the correct and AWS-aligned choice.

The Amazon SageMaker Model Registry is designed to manage and catalog ML models, including those hosted in Amazon ECR. By creating a model group for each model in the SageMaker Model Registry and setting up cross-account resource policies, the company can establish a central catalog in a new AWS account. This allows all models from the initial accounts to be accessible in a unified, centralized manner for better organization, management, and governance. This solution leverages existing AWS services and ensures scalability and minimal operational overhead.

Scenario: The company wants to perform online validation of a new ML model on 10% of the traffic before fully deploying the model in production. The setup must have minimal operational overhead.

Why Use SageMaker Production Variants?

Built-In Traffic Splitting: Amazon SageMaker endpoints support production variants, allowing multiple models to run on a single endpoint. You can direct a percentage of incoming traffic to each variant by adjusting the variant weights.

Ease of Management: Using production variants eliminates the need for additional infrastructure like separate endpoints or custom ALB configurations.

Monitoring with CloudWatch: SageMaker automatically integrates with CloudWatch, enabling real-time monitoring of model performance and invocation metrics.

Steps to Implement:

Deploy the New Model as a Production Variant:

Update the existing SageMaker endpoint to include the new model as a production variant. This can be done via the SageMaker console, CLI, or SDK.

Example SDK Code:

import boto3

sm_client = boto3.client( ' sagemaker ' )

response = sm_client.update_endpoint_weights_and_capacities(

EndpointName= ' existing-endpoint-name ' ,

DesiredWeightsAndCapacities=[

{ ' VariantName ' : ' current-model ' , ' DesiredWeight ' : 0.9},

{ ' VariantName ' : ' new-model ' , ' DesiredWeight ' : 0.1}

]

)

Set the Variant Weight:

Assign a weight of 0.1 to the new model and 0.9 to the existing model. This ensures 10% of traffic goes to the new model while the remaining 90% continues to use the current model.

Monitor the Performance:

Use Amazon CloudWatch metrics, such as InvocationCount and ModelLatency, to monitor the traffic and performance of each variant.

Validate the Results:

Analyze the performance of the new model based on metrics like accuracy, latency, and failure rates.

Why Not the Other Options?

Option B: Setting the weight to 1 directs all traffic to the new model, which does not meet the requirement of splitting traffic for validation.

Option C: Creating a new endpoint introduces additional operational overhead for traffic routing and monitoring, which is unnecessary given SageMaker ' s built-in production variant capability.

Option D: Configuring the ALB to route traffic requires manual setup and lacks SageMaker ' s seamless variant monitoring and traffic splitting features.

Conclusion:

Using production variants with a weight of 0.1 for the new model on the existing SageMaker endpoint provides the required traffic split for online validation with minimal operational overhead.

This is a regression problem, where the target variable is continuous and numeric. AWS documentation clearly states that classification metrics such as accuracy and AUC are not appropriate for regression models.

Root Mean Square Error (RMSE) measures the square root of the average squared differences between predicted and actual values. RMSE penalizes larger errors more heavily, making it especially useful when large prediction errors are costly or undesirable.

SageMaker Canvas automatically selects regression metrics such as RMSE and MAE when building regression models. RMSE is widely used for time-based and numeric prediction problems, especially when evaluating long historical datasets.

Inference latency measures system performance, not model accuracy.

Therefore, Option D is the correct and AWS-verified answer.

The temperature parameter controls the randomness in the model ' s responses. Lowering the temperature makes the model produce more deterministic and consistent answers.

The top_k parameter limits the number of tokens considered for generating the next word. Reducing top_k further constrains the model ' s options, ensuring more predictable responses.

By decreasing both parameters, the responses become more focused and consistent, reducing variability in similar queries.

Option A is correct because Amazon SageMaker Feature Store is the AWS service designed to act as a centralized repository for ML features that are used consistently across training and inference. AWS documentation states that SageMaker Feature Store simplifies how you create, store, share, and manage features for data exploration, model training, and model inference. This directly matches the requirement to create a repository for streamed song ratings that will be used in both batch training and real-time inference.

The most important requirement in the question is that the data must stay synchronized between batch training and real-time inference. AWS documents explain that Feature Store provides both an offline store and an online store. The offline store is used for historical data, model training, and batch inference, while the online store is a low-latency, high-availability store intended for real-time lookup during inference. This dual-store design is exactly why Feature Store is used to maintain feature consistency across training and serving workflows. AWS Well-Architected guidance also explicitly says Feature Store provides online storage for real-time inference and offline storage for model training and batch inference.

The other options do not solve the full problem. Athena CTAS can organize query results but does not provide a synchronized feature repository for online and offline ML use. Lake Formation governs access to data lakes but is not a feature repository for training and inference consistency. Data Wrangler Generate Data Insights is for analysis and preparation, not synchronized feature serving. Therefore, the best AWS-documented answer is A.

SageMaker script mode allows you to bring existing custom Python scripts and run them on AWS with minimal changes. SageMaker provides prebuilt containers for ML frameworks like PyTorch, simplifying the migration process. This approach enables the company to leverage their existing Python scripts and domain knowledge while benefiting from the scalability and managed environment of SageMaker. It requires the least effort compared to building custom containers or retraining models from scratch.

SageMaker Debugger can identify when a training job is not converging or is stuck in a non-productive state. By stopping these jobs early, unnecessary energy and computational resources are conserved, improving sustainability.

AWS Trainium instances are purpose-built for ML training and are optimized for energy efficiency and cost-effectiveness. They use less energy per training task compared to general-purpose instances, making them a sustainable choice.

The key requirement in this scenario is detecting customer intent without having any training data. According to AWS Machine Learning and Generative AI documentation, zero-shot learning is specifically designed for situations where labeled training data is unavailable. Zero-shot learning allows a pre-trained large language model (LLM) to perform tasks it has not been explicitly trained on by leveraging its general knowledge and language understanding.

Amazon Bedrock provides fully managed access to foundation models (FMs) and LLMs that support zero-shot and few-shot learning. By using an LLM from Amazon Bedrock, the company can directly infer customer intent from natural language inputs without building, training, or fine-tuning a custom model. This approach is ideal for conversational interfaces where rapid deployment and scalability are required.

Option A is incorrect because fine-tuning a sequence-to-sequence (seq2seq) model in Amazon SageMaker JumpStart still requires labeled training data. Since the company explicitly does not have training data, this option does not meet the requirement.

Option C is also incorrect because the Amazon Comprehend DetectEntities API is designed for named entity recognition (NER), such as detecting names, dates, locations, or monetary values. It does not perform intent detection and is not suitable for conversational AI intent classification.

Option D is partially misleading. While it is technically possible to run an LLM on Amazon EC2, this does not inherently solve the problem of intent detection without training data. Additionally, Amazon Bedrock already abstracts infrastructure management, scaling, and model hosting, making direct EC2 deployment unnecessary and less efficient.

Therefore, using an LLM from Amazon Bedrock with zero-shot learning is the most appropriate, scalable, and AWS-recommended solution for intent detection without training data.

The AWS::SageMaker::Model resource in AWS CloudFormation is used to create an ML model in Amazon SageMaker. This model can then be hosted on an endpoint by using the AWS::SageMaker::Endpoint resource. The model resource defines the container or algorithm to use for hosting and the S3 location of the model artifacts.

For real-time video content moderation with minimal operational overhead, AWS documentation recommends using fully managed, purpose-built AI services. Amazon Rekognition provides real-time video analysis capabilities, including content moderation, unsafe content detection, and label recognition for live video streams.

By integrating Rekognition with AWS Lambda, the company can automatically process video frames, extract moderation metadata, and take immediate action (such as flagging or stopping a stream) without managing servers, models, or infrastructure. This serverless architecture scales automatically and minimizes operational complexity.

Option B introduces unnecessary complexity. While Amazon Bedrock LLMs are powerful, they are not required for image-based moderation tasks that Rekognition already handles natively.

Option C is incorrect because using Amazon SageMaker would require model training, endpoint management, and scaling, significantly increasing operational overhead.

Option D is incorrect because Amazon Transcribe and Amazon Comprehend are designed for audio and text analysis, not image or video frame moderation.

Therefore, Amazon Rekognition with AWS Lambda is the most efficient, scalable, and low-maintenance solution for real-time video moderation during live streaming events.

The target_precision hyperparameter in the Amazon SageMaker linear learner controls the trade-off between precision and recall for the model. Increasing the target_precision prioritizes minimizing false positives by making the model more cautious in its predictions. This approach is effective for use cases where false positives have higher consequences than false negatives.

Option D is correct because Amazon Bedrock Knowledge Bases are designed for applications that need to answer questions using private documents without retraining or repeatedly fine-tuning a foundation model. AWS documentation states that with Amazon Bedrock Knowledge Bases, you can build applications enriched by context retrieved from a knowledge base, and that this provides an out-of-the-box RAG solution. AWS also explicitly says that adding a knowledge base increases cost-effectiveness by removing the need to continually train your model to use your private data. That matches this use case very closely.

The question also says the documentation consists of text files that are only several megabytes total and are updated often. A retrieval-based approach is more economical and operationally simpler than creating a new model or repeatedly fine-tuning one whenever the documents change. AWS documentation for Bedrock knowledge bases describes adding data sources and running ingestion jobs to process and index the content, which is exactly the pattern needed for frequently updated internal documentation used by a chat interface.

The other options are not as cost-effective. Creating a new LLM is far beyond the need here. Guardrails help control model behavior and policy enforcement, but they do not serve as a document retrieval layer for internal documentation. Fine-tuning a model on frequently changing text files is usually more expensive and less flexible than using retrieval augmentation. For a modest-sized, frequently updated documentation corpus, the AWS-native and most cost-effective solution is to load the files into an Amazon Bedrock knowledge base and use it to provide context at inference time. Therefore, the best verified answer is D.

A high max_depth value in XGBoost can lead to overfitting, where the model learns the training dataset too well but fails to generalize to new and unseen data. By decreasing the max_depth, the model becomes less complex, reducing overfitting and improving its ability to detect fraud in new transactions. This adjustment helps the model focus on general patterns rather than memorizing specific details in the training data.

The large gap between training accuracy (99%) and validation accuracy (82%) is a textbook case of overfitting. The model has learned patterns that fit the training data extremely well but do not generalize to unseen data.

AWS ML best practices recommend regularization techniques to address overfitting. Dropout layers randomly deactivate neurons during training, preventing the network from relying too heavily on specific paths. L1 and L2 regularization penalize large weights, reducing model complexity and improving generalization. k-fold cross-validation provides a more robust evaluation by training and validating the model across multiple data splits.

Option A increases complexity, which would worsen overfitting. Option C mixes valid ideas (dimensionality reduction) with unrelated changes (loss function choice) and is less targeted. Option D focuses on data quality but does not directly address model variance.

Therefore, implementing dropout, regularization, and k-fold cross-validation is the correct solution.

The correct answer is A. Create Amazon Managed Streaming for Apache Kafka (Amazon MSK) Serverless clusters to process the data.

Amazon MSK Serverless allows organizations to run Apache Kafka workloads without managing or provisioning the underlying infrastructure. Unlike MSK Provisioned clusters (Option B), serverless clusters automatically scale to accommodate variable workloads and handle operational tasks such as patching, scaling, and monitoring. This meets the company’s requirement to avoid managing or scaling infrastructure while providing near real-time streaming analytics.

Option C, using Amazon Kinesis Data Streams with Application Auto Scaling, provides a managed streaming solution but is not based on open-source technology. While Kinesis can handle near real-time data, the question specifically emphasizes the company’s desire for open-source technology, which points toward Kafka.

Option D, self-hosting Apache Flink on EC2, requires manual management of servers, scaling, patching, and container orchestration. This approach contradicts the requirement of not managing or scaling infrastructure and introduces operational complexity.

MSK Serverless provides seamless integration with open-source Kafka APIs, enabling applications to produce and consume streaming data in real time while AWS handles scaling and availability. It is ideal for analytics pipelines where data ingestion is continuous and variable, and infrastructure overhead must be minimized.

By choosing MSK Serverless, the company can implement near real-time updates for its analytics platform using open-source Kafka without the operational burden of cluster management, fully aligning with AWS best practices for deployment and orchestration of ML workflows in a managed, scalable, and serverless manner.

Amazon SageMaker Data Wrangler supports both direct and cataloged connections. To ensure data is always up to date, AWS recommends using cataloged connections backed by AWS Glue Data Catalog.

Cataloged connections allow Data Wrangler to reference the source systems dynamically, ensuring that each import reflects the latest data changes without manual reconfiguration. This approach supports Amazon S3, Amazon Redshift, and Snowflake and integrates securely using managed credentials.

Direct connections are point-in-time imports and do not automatically reflect schema or data updates. Glue and Lambda-based solutions introduce unnecessary complexity and operational overhead.

Therefore, using cataloged connections in Data Wrangler is the correct solution.

The text feature contains high-cardinality categorical values with minor spelling variations, which is a common real-world data quality issue. Traditional encoding techniques such as ordinal encoding (Option A) or one-hot encoding (Option C) would treat each misspelled value as a completely separate category, resulting in poor generalization and very high dimensionality.

Similarity encoding addresses this problem by grouping or encoding categories based on string similarity. Categories that differ only slightly—such as spelling errors—are encoded with similar numerical representations. Amazon SageMaker Data Wrangler supports similarity encoding specifically for such noisy categorical features.

Target encoding (Option D) depends on the target label distribution and risks target leakage if not carefully applied.

Therefore, similarity encoding is the most appropriate and AWS-recommended solution.

Option B is correct because AWS documentation states that speculative decoding is specifically a technique to speed up the decoding process of large language models. The question highlights that the model generates one token at a time and that users are unhappy with the time needed to generate lengthy responses. That is exactly the stage of inference targeted by speculative decoding. AWS explains that this method improves latency without compromising the quality of the generated text.

According to AWS, speculative decoding works by using a smaller, faster draft model to generate candidate tokens first. The larger target model then verifies those tokens. AWS further explains that the draft model can generate multiple candidate tokens quickly, and the target model evaluates them in parallel, which speeds up the final response. This directly addresses the problem in the scenario: the customer support agent is slow not because of startup time, but because long answers require many decoding steps.

The other options are not the best fit. Compilation reduces deployment time and auto-scaling latency through ahead-of-time compilation, but AWS documents it primarily as helping model deployment and hardware optimization rather than token-by-token response generation. Quantization reduces hardware requirements and can lower cost, but the docs do not present it as the most direct answer to slow sequential decoding. Fast model loading improves how quickly the model loads onto instances, which helps startup and scale-out time, not long response generation after the model is already serving traffic. Therefore, the best AWS-documented answer is B.

The requirement is to quickly determine whether the target variable is predictable, with minimal development effort. Amazon SageMaker Autopilot is specifically designed for this purpose.

SageMaker Autopilot automatically handles data preprocessing, feature engineering, algorithm selection, model training, and evaluation. It generates multiple candidate models and provides detailed performance metrics, allowing teams to quickly assess predictability without writing custom code.

Option B requires significant manual development. Option C is incorrect because Amazon Macie is a data security and classification service, not an ML modeling service. Option D is unsuitable because Amazon Bedrock models are not intended for structured tabular prediction tasks.

Therefore, SageMaker Autopilot provides the fastest and least effort solution.

For asynchronous inference on large datasets, AWS recommends SageMaker batch transform, which is designed for offline and large-scale inference workloads. Batch transform jobs do not require real-time endpoints and can process large volumes of data efficiently.

For monitoring data quality, AWS provides Amazon SageMaker Model Monitor, which supports scheduled monitoring of data quality, schema drift, and feature distribution drift. Model Monitor can publish metrics to Amazon CloudWatch and trigger alerts when thresholds are breached.

The other options lack native ML-specific monitoring capabilities. CloudTrail audits API activity, not data quality. Glue, Batch, and ECS would require extensive custom monitoring logic.

Therefore, SageMaker batch transform combined with Model Monitor is the correct solution.

The task described is unsupervised topic modeling, where topics are unknown in advance. Latent Dirichlet Allocation (LDA) is a probabilistic generative model specifically designed to discover latent topics from a corpus of documents without labeled data. AWS provides LDA as a built-in algorithm in Amazon SageMaker, making it well suited for this requirement.

LDA models documents as mixtures of topics and topics as mixtures of words, enabling interpretable topic discovery at scale. This aligns precisely with the need to organize documents into topics when the topics are not predefined.

BlazingText is optimized for word embeddings and supervised text classification, not topic modeling. Sequence-to-sequence models are used for translation or summarization. Object2Vec creates embeddings but does not itself perform topic discovery without additional clustering steps.

Therefore, LDA is the correct and purpose-built solution.

The requirement specifies real-time streaming ingestion and column-level transformation before sharing data with a vendor. Amazon Kinesis Data Streams is designed for low-latency, real-time data ingestion and delivery.

To extract only required columns from the stream, AWS recommends using Amazon Managed Service for Apache Flink as a stream consumer. Flink enables real-time transformations such as filtering, projection, and enrichment on streaming data before delivering it downstream.

Option A and D are batch-oriented and not suitable for real-time streaming. Option C is incorrect because S3 bucket policies cannot enforce column-level access controls.

Therefore, Kinesis Data Streams combined with Apache Flink meets all requirements.

Amazon SageMaker Clarify is the AWS service used to detect and quantify bias and fairness metrics in ML models. When bias monitoring jobs run, Clarify publishes bias metrics directly to Amazon CloudWatch.

CloudWatch metrics can be visualized using CloudWatch dashboards or integrated into other monitoring tools, making them ideal for real-time or periodic bias reporting.

CloudTrail logs API activity and does not capture ML metrics. EventBridge and SNS are used for event routing and notifications, not metric visualization.

AWS documentation explicitly states that Clarify bias metrics are emitted to Amazon CloudWatch, which is the correct source for dashboards.

Therefore, Option B is the correct and AWS-verified answer.

In AWS networking, security groups are stateful and allow-only, meaning they cannot explicitly deny traffic. As a result, Option A is invalid. Network ACLs (NACLs), on the other hand, are stateless and support both allow and deny rules, making them the correct mechanism for blocking traffic from specific IP addresses.

Because the SageMaker AI domain is deployed in a public subnet, inbound traffic reaches the subnet before it reaches the resource. AWS documentation states that NACLs are evaluated at the subnet level and are ideal for implementing IP-based blocking rules.

Route tables control routing paths, not traffic filtering, so Option D is incorrect. Option C is unrelated to network security and does not block traffic.

AWS best practices clearly recommend using network ACL deny rules when an explicit block is required for a specific IP address at the subnet boundary.

Therefore, Option B is the correct and AWS-aligned solution.

Large Parquet files can cause out-of-memory (OOM) issues during training if individual files exceed the memory capacity of the training instances. AWS documentation recommends repartitioning large datasets into smaller files to enable efficient streaming and parallel loading.

By using Apache Spark on Amazon EMR to repartition the Parquet files, the ML engineer can split the dataset into multiple smaller files that can be read incrementally during training. This approach avoids loading a single large file into memory and improves data parallelism.

Attaching larger EBS volumes increases storage capacity but does not solve memory constraints. Switching to memory-optimized instances increases cost and is not necessary when the dataset can be restructured. SageMaker distributed data parallelism focuses on model parameter synchronization across GPUs, not dataset file size.

AWS best practices explicitly recommend partitioning large Parquet datasets to improve memory efficiency during training.

Therefore, Option B is the correct and AWS-aligned solution.

Amazon SageMaker Automatic Model Tuning extracts objective metrics directly from training logs. For custom containers, AWS requires the ML engineer to define a metric definition that specifies a regular expression to parse the desired metric value from standard output.

The ObjectiveMetricName in the tuning job must match the metric captured by the regex. This is the only supported method for integrating custom metrics with SageMaker AMT.

TrainingInputMode does not define metrics. CloudWatch metrics cannot be used as tuning objectives. AMT cannot read metrics from S3 artifacts.

AWS documentation explicitly states that regex-based metric definitions are required for custom metrics in hyperparameter tuning jobs.

Therefore, Option B is the correct and AWS-verified solution.

Use case 1:

Summarize the text of a research paper

➡️ Sequence-to-Sequence (seq2seq) algorithm

Why:

Seq2seq models are designed for natural language generation tasks such as text summarization, translation, and paraphrasing. AWS documentation explicitly lists text summarization as a primary use case for the SageMaker seq2seq algorithm.

Use case 2:

Scan every pixel of an image to help self-driving cars identify objects in their path

➡️ Semantic segmentation algorithm

Why:

Semantic segmentation performs pixel-level classification, assigning a class label to every pixel in an image. This is exactly what is required for applications such as autonomous driving, road scene understanding, and object boundary detection.

Use case 3:

Identify abnormal data points in a dataset

➡️ Random Cut Forest (RCF) algorithm

Why:

Random Cut Forest is an unsupervised anomaly detection algorithm. AWS SageMaker RCF is purpose-built to identify outliers, unusual patterns, and anomalies in numerical datasets, making it ideal for fraud detection, monitoring, and abnormal data point detection.

AWS strongly recommends converting large CSV datasets into columnar formats such as Apache Parquet to improve query performance. Parquet reduces I/O by reading only the required columns and applies compression, which significantly speeds up analytics workloads.

AWS Glue ETL jobs provide a fully managed, serverless way to perform this conversion with minimal operational overhead. Once converted, the Parquet files can be queried efficiently by services such as Amazon Athena, Redshift Spectrum, and SageMaker processing jobs.

Splitting CSV files does not address inefficient storage format. Dropping columns risks data loss. Amazon EMR introduces infrastructure management overhead and is unnecessary for a straightforward format conversion.

AWS documentation clearly identifies CSV-to-Parquet conversion using Glue ETL as a best practice for scalable analytics.

Therefore, Option C is the correct answer.

In Amazon SageMaker AI, identifying bias in machine learning datasets before model training is a critical step to ensure fairness and reliability of predictions. This process is referred to as pre-training bias analysis, and it focuses on understanding whether the training data itself introduces bias—particularly through imbalanced class labels or sensitive attributes.

The Difference in Proportions of Labels (DPL) is a pre-training bias metric specifically designed to measure class imbalance. DPL compares the proportion of a specific label (such as a positive outcome) across different groups or classes within a dataset. If one class or group is overrepresented relative to another, the DPL value will deviate significantly from zero, clearly indicating imbalance. AWS documentation highlights DPL as a key metric used by SageMaker Clarify to detect label imbalance prior to model training.

By contrast, Mean Squared Error (MSE) is a regression evaluation metric used after model training to measure prediction error, not dataset bias. Silhouette score is an unsupervised learning metric used to evaluate clustering quality, making it irrelevant for supervised classification bias detection. Structural Similarity Index Measure (SSIM) is an image-quality metric used in computer vision tasks and has no application in dataset bias analysis.

Using DPL allows ML engineers to proactively detect and address skewed label distributions—such as by re-sampling, re-weighting, or collecting additional data—before training begins. This aligns with AWS best practices for responsible AI and helps reduce the risk of biased predictions that could negatively impact real-world decision-making.

Therefore, Difference in Proportions of Labels (DPL) is the correct and AWS-recommended metric for confirming class imbalance during pre-training bias analysis in Amazon SageMaker AI.

Predicting apartment prices is a regression problem, where the target variable is continuous. AWS documentation states that classification metrics such as accuracy, AUC, and F1 score are not appropriate for regression tasks.

Mean Absolute Error (MAE) measures the average absolute difference between predicted values and actual values. MAE is easy to interpret because it is expressed in the same units as the target variable (for example, dollars), making it especially useful for business-facing problems like price prediction.

AWS best practices recommend MAE for evaluating regression models when understanding average prediction error magnitude is important and when robustness to outliers is desired.

Therefore, Option D is the correct and AWS-aligned answer.