An ML engineer wants to use, prepare, and load data from Amazon S3 for analytics. The ML engineer must run an extract, transform, and load (ETL) job to discover the schema of the data and to store the metadata.
Which solution will meet these requirements with the LEAST manual effort?
A. Use AWS Glue to run the ETL job. Use the job to discover the schema and to store the associated metadata in the AWS Glue Data Catalog.
B. Create an Amazon SageMaker Data Wrangler flow to run the ETL job. Use the job to discover the schema and to store the associated metadata in an S3 bucket.
C. Create an ETL pipeline by using Amazon Athena integrated with AWs Step Functions. Use the pipeline to run the ETL job to discover the schema and to store the associated metadata in an S3 bucket.
D. Launch an Amazon EC2 instance that includes the scikit-learn library to run the ETL job. Use the job to discover the schema and to store the associated metadata in Amazon Redshift.
Show Answer
Correct Answer: A
Explanation: AWS Glue is the managed ETL service designed to crawl data in Amazon S3, automatically discover schemas with Glue Crawlers, run ETL jobs, and store table metadata in the AWS Glue Data Catalog. This requires the least manual effort compared with building custom pipelines or using services not intended for metadata cataloging.
Question 92
A company’s ML engineer is creating a classification model. The ML engineer explores the dataset and notices a column that is named day_of_week. The column’s data consists of the following values: Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, and Sunday.
Which technique should the ML engineer use to convert this column’s data to binary values?
A. Binary encoding
B. Label encoding
C. One-hot encoding
D. Tokenization
Show Answer
Correct Answer: C
Explanation: The day_of_week feature is a nominal categorical variable with seven distinct categories and no meaningful numeric ordering. One-hot encoding converts each category into its own binary indicator column, which is the standard approach. Label encoding would incorrectly impose an ordinal relationship, binary encoding is a different categorical encoding technique but not the standard binary indicator representation asked here, and tokenization is used for text processing.
Question 93
A company uses an Amazon SageMaker AI ML model to make real-time inferences. The company has configured auto scaling for the Amazon EC2 instances that SageMaker AI uses for the inferences.
During times of peak usage, new instances launch before existing instances are fully ready. As a result, the model experiences inefficiencies and delays.
Which solution will optimize the scaling process without affecting response times?
A. Change to a multi-model endpoint configuration in SageMaker AI.
B. Integrate Amazon API Gateway and AWS Lambda to manage invocations of the SageMaker AI inference endpoint.
C. Decrease the cooldown period for scale-in activities. Increase the maximum number of instances.
D. Increase the cooldown period after scale-out activities.
Show Answer
Correct Answer: D
Explanation: Increasing the scale-out cooldown period gives newly launched SageMaker endpoint instances time to finish initialization and begin serving inference traffic before additional scaling decisions are made. This reduces repeated premature scale-out events and improves scaling efficiency without increasing inference latency. The other options do not address the instance warm-up behavior that is causing the inefficiencies.
Question 94
A medical company ingests streams of data from devices that monitor patients’ vital signs. The company uses Amazon SageMaker and plans to prepare ML models to predict adverse events for patients. The dataset is large with thousands of features.
An ML engineer needs to run several hundred training iterations with different sets of features, different algorithms, and many potential parameters. The ML engineer must implement a solution to log the characteristics and results of each training iteration.
Which solution will meet these requirements with the LEAST implementation effort?
A. Use Amazon CloudWatch to create custom metrics for the characteristics of each iteration.
B. Write the characteristics of each iteration to logs in Amazon S3. Use AWS Glue and Amazon Athena to search the logs.
C. Use the SageMaker Model Registry to track the characteristics and results of each iteration.
D. Use SageMaker Experiments to track the characteristics and results of each iteration.
Show Answer
Correct Answer: D
Explanation: SageMaker Experiments is designed to organize and track ML experiments, including training runs, hyperparameters, feature sets, algorithms, input artifacts, and evaluation metrics. It enables comparison across hundreds of iterations with minimal implementation effort. CloudWatch is for monitoring metrics, S3/Glue/Athena requires custom logging and querying, and Model Registry is intended for managing and versioning models after training rather than tracking experiment iterations.
Question 95
A company is training a large language model (LLM) by using on-premises infrastructure. A live conversational engine uses the LLM to help customers find real-time insights in credit card data.
An ML engineer must implement a solution to train and deploy the LLM on Amazon SageMaker.
Which solution will meet these requirements?
A. Use SageMaker Training Compiler to train the LLM. Deploy the LLM by using SageMaker real-time inference.
B. Use SageMaker with deep learning containers for large model inference to train the LLM. Deploy the LLM by using SageMaker real-time inference.
C. Use SageMaker Notebook Jobs to train the LLM. Deploy the LLM by using SageMaker Asynchronous Inference.
D. Use SageMaker Studio to train the LLM. Deploy the LLM by using SageMaker batch transform.
Show Answer
Correct Answer: A
Explanation: SageMaker Training Compiler is designed to optimize and accelerate training of deep learning models, including supported transformer workloads, while SageMaker real-time inference is the appropriate deployment option for a live conversational application that requires low-latency responses. The Large Model Inference (LMI) containers are intended for inference hosting rather than training. Notebook Jobs and Studio are development/orchestration environments, not the primary training mechanism, and asynchronous inference or batch transform do not fit a real-time conversational workload.
Sources:
https://officialdumps.com/exam/mla-c01
Question 96
A company has several teams that have developed separate prediction models on their own laptops. The teams developed the models by using Python with scikit-learn and TensorFlow frameworks.
The company must rebuild the models and must integrate the models into an ML infrastructure that the company manages by using Amazon SageMaker. The company also must incorporate the models into a model registry.
Which solution will meet these requirements with the LEAST operational overhead?
A. Export the models from the laptops to an Amazon S3 bucket. Use an Amazon API Gateway REST API and AWS Lambda functions with SageMaker endpoints to access the models. Register the models in the SageMaker Model Registry.
B. Import the models into the SageMaker Model Registry. Use SageMaker to run the imported models.
C. Use code from the laptops to create containers for the models. Use the bring your own container (BYOC) functionality of SageMaker to import and use the models. Register the models in the SageMaker Model Registry.
D. Import the Python-based models into SageMaker. Rebuild the scikit-learn and TensorFlow models in SageMaker. Register all the models in the SageMaker Model Registry.
Show Answer
Correct Answer: D
Explanation: SageMaker provides built-in training and inference containers for scikit-learn and TensorFlow, so the models can be rebuilt in a managed SageMaker environment without maintaining custom containers. After training, the models can be registered directly in SageMaker Model Registry. Using BYOC adds container management overhead, and simply importing arbitrary locally trained models into the registry is not a supported generic workflow for rebuilding and managing them in SageMaker. API Gateway and Lambda are unrelated to the requirement to rebuild and register the models.
Question 97
HOTSPOT
-
An ML engineer must choose the appropriate Amazon SageMaker algorithm to solve specific AI problems.
Select the correct SageMaker built-in algorithm from the following list for each use case. Each algorithm should be selected one time.
• Random Cut Forest (RCF) algorithm
• Semantic segmentation algorithm
• Sequence-to-Sequence (seq2seq) algorithm
Show Answer
Correct Answer: Summarize research paper → Sequence-to-Sequence (seq2seq) algorithm
Scan every pixel for self-driving object identification → Semantic segmentation algorithm
Identify abnormal data points → Random Cut Forest (RCF) algorithm
Explanation: Seq2Seq is used for text generation tasks such as summarization. Semantic segmentation classifies each image pixel. Random Cut Forest is used for anomaly detection.
Question 98
An ML engineer wants to use a set of survey responses as training data for an ML classifier. All the survey responses are either “yes” or “no.”
The ML engineer needs to convert the responses into a feature that will produce better model training results. The ML engineer must not increase the dimensionality of the dataset.
Which methods will meet these requirements? (Choose two.)
A. Binary encoding
B. Label encoding
C. One-hot encoding
D. Statistical imputation
E. Tokenization
Show Answer
Correct Answer: A, B
Explanation: For a binary categorical feature with values 'yes' and 'no', label encoding maps the categories to a single numeric column (for example, yes=1, no=0) without increasing dimensionality. Binary encoding also represents categorical values using binary digits and, for a two-category feature, does not increase the number of dimensions in practice. One-hot encoding adds columns, statistical imputation is for missing values, and tokenization is for splitting text rather than encoding a binary categorical feature.
Question 100
A company is developing a new ML model that uses the XGBoost algorithm. The company will train the model on data that is stored in an Amazon S3 bucket. The data is in a nested JSON format.
An ML engineer needs to convert the JSON files into a tabular format.
Which solution will meet this requirement with the LEAST operational overhead?
A. Create an AWS Glue PySpark job that uses the Relationalize transform to convert the files.
B. Write custom Scala code to convert the files. Use Amazon EMR Serverless to run the Scala code.
C. Create an AWS Lambda function that uses a Python runtime and invokes the reduce() function to convert the files. Invoke the Lambda function.
D. Create an Amazon Athena database that is based on the JSON files. Use the Athena flatten function to convert the data.
Show Answer
Correct Answer: A
Explanation: AWS Glue provides a managed ETL service with a built-in Relationalize transform specifically designed to flatten nested JSON into tabular structures. This minimizes custom code and operational overhead compared to managing EMR jobs, implementing custom Lambda transformations, or relying on Athena, which is primarily a query service and not intended for transforming datasets into tabular training data.
Question 101
A company is building an ML model by using Amazon SageMaker, AWS owned libraries, and open source libraries. The company must ensure that SageMaker does not collect metadata about usage and errors during training.
Which solution will meet these requirements?
A. Associate the SageMaker domain with a custom IAM role. Attach the role to a policy that denies Amazon CloudWatch service usage logs.
B. Add an IAM role to the SageMaker domain to deny Amazon CloudWatch the permission to report metadata.
C. Turn off the setting in the SageMaker domain to share metadata for console jobs. Opt out of metadata collection for each training job that is submitted through the AWS CLI or AWS SDKs.
D. Set a parameter to opt out of metadata collection for each training job that is submitted through the AWS CLI, Boto3, or the SageMaker Python SDK.
Show Answer
Correct Answer: D
Explanation: Amazon SageMaker provides an opt-out parameter for metadata collection on training jobs submitted through supported SDKs and the AWS CLI. IAM policies denying CloudWatch access do not disable SageMaker's metadata collection, and the console sharing setting applies to console jobs rather than all programmatic training jobs. Therefore, setting the opt-out parameter for each training job is the required solution.
$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.
