MLS-C01 Practice Test 44 — 15 Questions

Question 1

A company is training a deep learning model on a large dataset using Amazon SageMaker. The training script uses TensorFlow and requires GPUs. The training job is failing with an out-of-memory error. Which configuration change should be made to resolve this issue?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Use a larger instance type with more GPU memory.. The training job is failing with an out-of-memory error, which indicates that the model or batch size exceeds the GPU memory capacity of the current instance. Using a larger instance type with more GPU memory directly addresses this by providing additional VRAM, allowing the model to fit in memory and the training to proceed without failure.

Answer

Increase the number of instances in the training job.

Answer

Switch to using spot instances to reduce cost.

Answer

Enable distributed training across multiple instances.

Question 2

A company is using Amazon SageMaker to train a model. The training data includes sensitive personally identifiable information (PII). The company needs to ensure that the training data is protected and that the trained model does not inadvertently expose PII. Which TWO actions should the company take? (Choose TWO.)

Accepted Answer

Encrypt the training data in S3 using AWS KMS. Option A is correct because encrypting the training data in Amazon S3 with AWS KMS (SSE-KMS) protects the PII at rest, ensuring that even if storage is compromised the data remains unreadable without the appropriate KMS key permissions. Option C is correct because using SageMaker's data processing (e.g., SageMaker Processing jobs with PII detection/redaction logic) to redact or anonymize PII before training prevents the model from learning and later exposing sensitive information. Option B is not sufficient because SSE-S3 uses S3-managed keys, which offer less control, auditing, and customer-managed key separation than AWS KMS for sensitive PII. Option D does not protect data or prevent PII exposure; CloudTrail only provides audit logging of API activity. Option E is wrong because granting public read access would directly expose the PII and violate the protection requirement.

Answer

Use server-side encryption with S3-managed keys

Answer

Enable AWS CloudTrail to log all access to the data

Answer

Grant public read access to the training data for faster access

Question 3

A machine learning team is building a real-time inference pipeline using Amazon SageMaker. The input data is located in an S3 bucket, and the team needs to transform the data before inference using a custom Python script. The transformation should run on a serverless infrastructure and must be triggered automatically when new data arrives in S3. Which combination of services should the team use?

Accepted Answer

Use AWS Lambda functions triggered by S3 events to run the transformation, then invoke a SageMaker endpoint.. AWS Lambda functions can be triggered directly by S3 events (e.g., ObjectCreated) to run a custom Python transformation script on the incoming data, and then invoke a SageMaker endpoint for real-time inference. This combination meets the serverless infrastructure requirement and provides automatic, event-driven processing without managing any servers.

Answer

Use AWS Glue jobs triggered by S3 events.

Answer

Use Amazon SageMaker Processing jobs triggered by S3 events.

Answer

Use Amazon Kinesis Data Firehose to transform data and deliver to SageMaker.

Question 4

A company is using Amazon SageMaker to deploy a machine learning model that predicts equipment failure. The model is a binary classifier that outputs a probability. The company wants to set a threshold such that the model correctly identifies 95% of actual failures (recall >= 0.95). The model's precision at the current threshold of 0.5 is 0.7. The data scientist evaluates the model on a test set and obtains the following confusion matrix at threshold 0.5: TP=95, FN=5, FP=40, TN=860. The total actual positives are 100. Which threshold adjustment should the data scientist make to achieve the recall goal?

Accepted Answer

Keep the threshold at 0.5. The current threshold of 0.5 already yields TP=95 and FN=5, giving recall = 95/(95+5) = 0.95, which meets the stated goal of recall >= 0.95. No threshold adjustment is needed. Decreasing the threshold would increase recall further but also increase false positives, while increasing the threshold would reduce recall below the target.

Answer

Decrease the threshold to 0.1

Answer

Increase the threshold to 0.7

Answer

Decrease the threshold to 0.3

Question 5

A machine learning team is building a fraud detection model. The dataset is highly imbalanced (99.9% legitimate, 0.1% fraudulent). Which EDA technique is most important to apply before modeling?

Accepted Answer

Create a stratified train-test split to preserve the class distribution.. With a 99.9%/0.1% class imbalance, a random train-test split can easily produce a test set with zero or very few fraud examples, making evaluation meaningless. A stratified split preserves the class ratio in both training and test sets, ensuring the minority class is represented proportionally so metrics like precision, recall, and AUC-PR are meaningful. This is the most critical EDA/preprocessing step before modeling an imbalanced fraud dataset.

Answer

Normalize all numerical features to have zero mean and unit variance.

Answer

Remove outliers from the dataset using the IQR method.

Answer

Perform correlation analysis to remove highly correlated features.

Question 6

A data scientist is building a recommendation system for an e-commerce platform. The dataset contains user interactions (clicks, purchases) and item metadata. The scientist wants to use matrix factorization. Which algorithm should be used?

Accepted Answer

SageMaker Factorization Machines. SageMaker Factorization Machines is specifically designed for recommendation systems and matrix factorization tasks. Option A (SageMaker Image Classification) is used for image classification, not matrix factorization. Option B (SageMaker BlazingText) is used for text classification or word embeddings, not for recommendation. Option C (SageMaker XGBoost) is a gradient boosting algorithm for regression and classification, not matrix factorization.

Answer

SageMaker Image Classification

Answer

SageMaker BlazingText

Answer

SageMaker XGBoost

Question 7

A data scientist is training a neural network on a GPU instance in Amazon SageMaker. The training job fails with an 'OutOfMemoryError'. Which action should the data scientist take to resolve this issue?

Accepted Answer

Reduce the batch size in the training script.. An OutOfMemoryError during GPU training indicates that the GPU's memory is exhausted. Reducing the batch size directly decreases the memory footprint per training step, as fewer samples and their corresponding activations are stored simultaneously. This is the most immediate and effective fix without changing the instance type or training architecture.

Answer

Enable automatic hyperparameter tuning.

Answer

Switch to distributed training across multiple instances.

Answer

Use a smaller instance type with less GPU memory.

Question 8

A company is using Amazon SageMaker to train a large language model. The training job is taking too long. The data scientist wants to reduce training time without sacrificing model accuracy. Which THREE strategies are MOST appropriate?

Accepted Answer

Use mixed precision training (float16). Mixed precision training (float16) reduces memory usage and accelerates computation by using half-precision floating-point numbers for most operations, while maintaining a single-precision copy of critical parameters to preserve accuracy. This directly reduces training time on compatible GPUs (e.g., NVIDIA V100, A100) without sacrificing model quality, as the loss scaling technique prevents underflow in gradients.

Answer

Switch from GPU instance to CPU instance

Answer

Increase the maximum sequence length

Question 9

A data scientist needs to analyze a dataset stored in Amazon S3 as CSV files. The dataset contains 100 columns, and the data scientist wants to quickly understand the distribution of each column, including missing values, data types, and basic statistics. Which AWS service is best suited for this task?

Accepted Answer

AWS Glue DataBrew. AWS Glue DataBrew is a visual data preparation service that profiles datasets and surfaces column statistics, data types, missing values, and distributions without writing code. It natively connects to S3 CSV sources and is designed for exactly this exploratory profiling task. This makes it the best fit for quickly understanding a 100-column dataset.

Answer

Amazon SageMaker Data Wrangler

Answer

Amazon QuickSight

Answer

Amazon Athena

Question 10

A data engineer is designing a streaming pipeline using Amazon Kinesis Data Analytics for Apache Flink. The pipeline reads from a Kinesis data stream and writes to a S3 bucket. The job must recover quickly from failures without reprocessing large amounts of data. Which TWO configurations should be used? (Choose TWO)

Accepted Answer

Enable checkpointing with a state backend like RocksDB.. Option A is correct because enabling checkpointing with a durable state backend such as RocksDB (via the EmbeddedRocksDBStateBackend or RocksDBStateBackend) allows Apache Flink to periodically snapshot operator state, so after a failure the job resumes from the last completed checkpoint instead of replaying the entire stream, which directly satisfies the requirement to recover quickly without reprocessing large amounts of data. Option C is correct because configuring the S3 sink with exactly-once delivery semantics (for example, using the Flink FileSink with RollingPolicy and OnCheckpointRollingPolicy, or the StreamingFileSink with exactly-once mode) ties file commit to completed checkpoints, ensuring that on recovery only data after the last checkpoint is reprocessed and no duplicates or partial files are written to S3. Option B is not appropriate because an in-memory state backend keeps state only in the TaskManager heap and does not persist snapshots durably, so it cannot support fast recovery from failures. Option D is incorrect because setting parallelism to the maximum number of shards is a throughput/scaling consideration and does not by itself provide fast failure recovery or avoid reprocessing. Option E is incorrect because extending Kinesis stream retention to 365 days only lengthens how long records remain readable; it does not improve checkpoint-based recovery or reduce reprocessing after a job failure.

Answer

Use in-memory state backend for low latency.

Answer

Set the parallelism to the maximum number of shards.

Answer

Increase the retention period of the Kinesis stream to 365 days.

Question 11

A data scientist is using Amazon SageMaker to train a model, but the training job fails with an 'Out of memory' error. The instance type is ml.p3.2xlarge. Which action should the data scientist take to resolve the issue?

Accepted Answer

Reduce the mini-batch size in the training script.. The 'Out of memory' error on a single ml.p3.2xlarge instance indicates that the GPU memory is insufficient for the current workload. Reducing the mini-batch size directly decreases the memory footprint per training step, allowing the model to fit within the available GPU memory without changing the instance type or incurring additional costs.

Answer

Use Pipe input mode.

Answer

Increase the number of instances.

Answer

Use a Spot instance.

Question 12

A company wants to build a data lake on Amazon S3. The data lake should support both batch and real-time data ingestion. Which AWS services should be used for data ingestion? (Choose TWO.)

Accepted Answer

AWS Glue. AWS Glue is correct because it provides a managed ETL service that can handle batch data ingestion into a data lake on Amazon S3. It can be scheduled for periodic batch loads or triggered by events, making it suitable for batch ingestion workflows. Amazon Kinesis Data Firehose is correct because it is a fully managed service for loading streaming data into S3 in near real-time, supporting real-time ingestion with automatic buffering and compression.

Answer

Amazon Redshift

Answer

Amazon Athena

Answer

Amazon SQS

Question 13

A machine learning team is using SageMaker to train a model. The training data is stored in an S3 bucket encrypted with AWS KMS. The training job fails with an 'AccessDenied' error. Which IAM permission is MOST likely missing from the SageMaker execution role?

Accepted Answer

kms:Decrypt. The training data is stored in an S3 bucket encrypted with AWS KMS. When SageMaker reads the encrypted data, the execution role must have permission to decrypt the KMS key. Without `kms:Decrypt`, the role cannot access the encrypted objects, resulting in an 'AccessDenied' error even if S3 read permissions are present.

Answer

s3:GetObject

Answer

s3:ListBucket

Answer

kms:GenerateDataKey

Question 14

Refer to the exhibit. An ML engineer attaches this IAM policy to a user. The user wants to invoke the SageMaker endpoint my-endpoint from an EC2 instance with public IP 52.1.1.1. What will happen?

Accepted Answer

The invocation is allowed because the source IP is not in the denied ranges.. The IAM policy explicitly allows the `sagemaker:InvokeEndpoint` action, and the `Deny` statement only denies requests from IP addresses in the ranges 10.0.0.0/8 or 192.168.0.0/16. Since the EC2 instance has a public IP of 52.1.1.1, which is not within those denied ranges, the invocation is allowed. The policy does not require the user to be in a VPC or to have endpoint creation permissions for invoking an existing endpoint.

Answer

The invocation fails because the user does not have permission to create an endpoint.

Answer

The invocation is denied because the Deny statement applies to all resources.

Answer

The invocation is denied because the user is not in a VPC.

Question 15

A data scientist is deploying a model on Amazon SageMaker. The model requires inference on images, and the data scientist wants to use a GPU instance for low latency. However, the data scientist is unsure about the instance type to choose for the endpoint. Which TWO factors should the data scientist consider when selecting the instance type? (Choose TWO.)

Accepted Answer

The cost per inference for the instance type. When selecting an instance type for a SageMaker endpoint, the data scientist should consider the cost per inference for the instance type (option C), because endpoint instances are billed continuously while running, so the instance's hourly price directly determines the cost efficiency of serving predictions, especially for low-latency GPU workloads. The data scientist should also consider the GPU memory available on the instance (option E), since image inference models (e.g., CNN-based models) must fit their model weights and intermediate tensors into GPU memory; insufficient GPU memory causes out-of-memory errors or forces smaller batch sizes that hurt latency. Option A is not relevant because training time is determined during the training job, not by the inference endpoint instance type. Option B is not the key factor here because the workload is GPU-based, so GPU memory and GPU compute matter more than vCPU count for low-latency image inference. Option D is irrelevant because the AWS Region of the S3 bucket storing the model does not determine the endpoint instance type; the model artifacts are downloaded to the endpoint instance regardless of bucket region.

Answer

The time taken to train the model

Answer

The number of vCPUs on the instance

Answer

The AWS Region of the S3 bucket storing the model