MLS-C01 Mock Exam 71 — 25 Questions

Question 1

A data scientist is using Amazon SageMaker to train a model with a built-in algorithm. They want to monitor the training job's resource utilization, such as CPU and GPU utilization, in real-time. Which AWS service can they use to view these metrics?

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

Accepted Answer

Amazon CloudWatch. Amazon CloudWatch is the AWS service that collects and displays metrics for SageMaker training jobs, including CPU and GPU utilization. It provides real-time monitoring and can trigger alarms. Other services like CloudTrail, Debugger, and X-Ray serve different purposes and do not offer the same resource utilization metrics.

Answer

AWS X-Ray

Answer

Amazon SageMaker Debugger

Answer

AWS CloudTrail

Question 2

A data scientist is analyzing a dataset with a highly imbalanced binary target variable (1% positive class). The scientist wants to perform exploratory data analysis to understand feature distributions and relationships with the target. The dataset is stored in Amazon S3 and is too large to fit into memory on a single Amazon SageMaker notebook instance. Which approach is most efficient for performing EDA on this dataset?

Accepted Answer

Use Amazon SageMaker Data Wrangler to import the dataset, then use its built-in analyses to visualize feature distributions and correlations with the target, leveraging its ability to sample large datasets.. SageMaker Data Wrangler is the most efficient because it is purpose-built for EDA on large datasets. It can sample data, handle imbalanced classes, and provide built-in visualizations without requiring custom code or downloading data. Other options either lack integration, require manual coding, or are impractical for large datasets.

Answer

Use Amazon Athena to run SQL queries that compute summary statistics and histograms, then visualize the results in Amazon QuickSight. This requires no additional setup and is the fastest method.

Answer

Use Amazon SageMaker Processing with a scikit-learn script to compute statistics and generate plots, storing the results in S3. This allows full control over the analysis but requires writing custom code.

Answer

Download the dataset to a local machine with more memory, then use pandas and matplotlib to perform EDA. This avoids the memory limitations of the SageMaker notebook instance.

Question 3

A data scientist is exploring a dataset containing a categorical feature with 50 unique values. The scientist wants to understand the relationship between this categorical feature and a continuous target variable. The dataset is stored in Amazon S3 and is moderately large (10 GB). Which visualization approach is most suitable for this exploratory analysis?

Accepted Answer

Create a box plot for each category of the categorical feature, showing the distribution of the target variable within each category.. A box plot is the most suitable because it displays the distribution of the continuous target variable for each category, allowing comparison across the 50 categories. It handles multiple categories well and reveals medians, spread, and outliers. Other visualization types are either inappropriate for categorical data or fail to show the distribution effectively.

Answer

Create a line chart with the categorical feature on the x-axis and the mean target value on the y-axis, connecting the points to show trends.

Answer

Create a heatmap of the correlation matrix between the categorical feature and the target variable after one-hot encoding the categorical feature.

Answer

Create a scatter plot of the categorical feature versus the target variable, using jitter to avoid overplotting.

Question 4

A data scientist is using Amazon SageMaker to train a model with a built-in algorithm. They want to monitor the training job's resource utilization, such as CPU and GPU utilization, in real-time. Which AWS service can they use to view these metrics?

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

Accepted Answer

Adjust the scale_pos_weight parameter in the XGBoost hyperparameters to a higher value.. For imbalanced binary classification with XGBoost, the scale_pos_weight hyperparameter is designed to handle class imbalance by weighting the positive class more heavily. Setting it to the ratio of negative to positive instances helps the model pay more attention to the minority class, improving recall. Other methods like resampling or using evaluation metrics like AUC-PR can also help, but scale_pos_weight is a direct and effective hyperparameter adjustment.

Answer

Apply feature selection to remove noisy features that may be causing the model to misclassify positives.

Answer

Increase the number of boosting rounds to allow the model to learn the positive class better.

Answer

Use a different evaluation metric such as accuracy instead of recall to guide hyperparameter tuning.

Question 5

A data scientist is using Amazon SageMaker to train a model with a built-in algorithm. They want to monitor the training job's resource utilization, such as CPU and GPU utilization, in real-time. Which AWS service can they use to view these metrics?

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

Accepted Answer

Create an IAM role with a policy that allows s3:GetObject on the bucket and pass it as the execution role to the training job. SageMaker training jobs use an IAM execution role to access AWS resources. Creating a role with the necessary S3 read permissions and specifying it when launching the training job is the simplest and most secure method. Other options either do not apply to the training job's context or pose security risks.

Answer

Make the S3 bucket public so that the training job can access the data without credentials

Answer

Create an IAM role with the AmazonS3ReadOnlyAccess policy and attach it to the SageMaker notebook instance

Answer

Use the AWS root user credentials to create the training job

Question 6

A machine learning engineer is tuning a SageMaker XGBoost model. They configure a hyperparameter tuning job with a maximum of 30 training jobs and set the objective metric as 'validation:auc'. After the tuning job completes, they want to deploy the model that achieved the best validation AUC. Which step will identify the best training job?

Accepted Answer

Call DescribeHyperParameterTuningJob and use the BestTrainingJob field.. After a hyperparameter tuning job completes, the BestTrainingJob field in DescribeHyperParameterTuningJob provides the name of the training job that achieved the best objective metric. Deploying that model ensures you use the optimal hyperparameters found during tuning. Other methods do not reliably identify the best model.

Answer

Use the SageMaker console to sort the tuning job's training jobs by creation date and pick the latest.

Answer

List all training jobs with ListTrainingJobs and select the one with the shortest training time.

Answer

Retrieve the best model from the tuning job's S3 output path by reading the manifest file.

Question 7

A data scientist is exploring a dataset of customer transactions stored in Amazon S3 as Parquet files. The dataset contains a numerical feature 'transaction_amount' with many outliers, and the scientist wants a robust measure of central tendency and a robust measure of spread that are not influenced by extreme values. Which pair of statistics should the scientist compute?

Accepted Answer

Median and interquartile range (IQR). The median and interquartile range are robust statistics because they rely on the order of values rather than the actual magnitudes, so extreme transaction amounts do not distort them. In exploratory data analysis, when a numerical feature has outliers, these measures give a more representative summary than mean and standard deviation or variance.

Answer

Mean and variance

Answer

Mean and standard deviation

Answer

Mode and range

Question 8

A machine learning engineer is using Amazon SageMaker Automatic Model Tuning to optimize a gradient boosting model. The tuning job has completed, and the engineer wants to identify the best hyperparameter configuration. Which of the following should the engineer do to find the best configuration?

Accepted Answer

Examine the DescribeHyperParameterTuningJob response and look for the BestTrainingJob field.. After a hyperparameter tuning job completes, the best training job is automatically identified based on the objective metric. The DescribeHyperParameterTuningJob API returns this information in the BestTrainingJob field, which includes the hyperparameters and the metric value. This is the most reliable and direct way to obtain the best configuration.

Answer

Query the SageMaker Experiments API to retrieve the trial component with the highest metric value.

Answer

Sort the training jobs by the objective metric in Amazon CloudWatch Logs.

Answer

Use the SageMaker console to view the tuning job and manually compare the hyperparameters of each training job.

Question 9

A machine learning engineer is preparing a dataset for a binary classification model. The dataset contains 50 numeric features and one binary target. During exploratory data analysis, the engineer wants to identify which features are most strongly associated with the target, but suspects that some features have non-linear relationships with the target and that a few features are highly correlated with each other. Which approach is most appropriate to quantify feature-target associations while accounting for non-linearity and also detect multicollinearity?

Accepted Answer

Compute the mutual information between each feature and the target, and compute the variance inflation factor (VIF) for each feature.. Mutual information is a non-parametric measure that captures any dependency between a feature and the target, including non-linear patterns, making it ideal for the stated suspicion. Variance inflation factor quantifies multicollinearity by measuring how much the variance of a coefficient is inflated due to correlation with other features. Using both together provides a comprehensive EDA view without needing to train a model.

Answer

Compute the Pearson correlation coefficient between each feature and the target, and compute the variance inflation factor (VIF) for each feature.

Answer

Train a random forest classifier and extract feature importances, then compute pairwise Pearson correlations among all features to detect multicollinearity.

Answer

Perform principal component analysis (PCA) on the features and retain components with eigenvalues greater than 1, then interpret the loadings to identify important features.

Question 10

A company runs an AWS Glue extract, transform, and load job that reads semi-structured JSON from Amazon S3, flattens nested arrays, and writes Parquet to a curated bucket. The job has grown to 200 DPU-hours and fails intermittently with out-of-memory errors on executors. A data engineer must reduce runtime and memory pressure without changing the output schema. (Choose two.)

Accepted Answer

Enable the Glue job bookmark and process only new S3 objects on each run.. Job bookmarks restrict each run to newly arrived objects, shrinking the data volume that flows through the pipeline, while additional workers with auto scaling spread partitions across more executors to lower per-executor memory use and shorten runtime. Both changes affect throughput and resource distribution only; they leave the transformation logic and the curated Parquet schema exactly as before.

Answer

Convert the job to use the AWS Glue Python shell job type instead of Spark.

Answer

Apply a ResolveChoice transform to cast all nested fields to string before the flatten step.

Answer

Set the job's maximum concurrency to a high value and schedule overlapping runs.

Question 11

A company uses Amazon SageMaker to train a deep learning model for image classification. The training job is using a single p3.2xlarge instance and is taking too long. The dataset is stored in Amazon S3 and consists of 1 million images. The company wants to reduce training time without changing the model architecture. Which of the following should be done to achieve this?

Accepted Answer

Enable SageMaker distributed training using the Horovod framework and multiple instances.. Distributed training with Horovod across multiple instances parallelizes the training process, directly reducing the time required for each epoch. Since the model architecture is fixed, data parallelism is appropriate. Other options like Pipe mode or volume size may offer minor improvements but do not address the core compute bottleneck. Converting to RecordIO with File mode does not add parallelism.

Answer

Convert the dataset to the RecordIO format and use File mode.

Answer

Increase the size of the training volume attached to the p3.2xlarge instance.

Answer

Use SageMaker Pipe mode instead of File mode to stream data directly from S3 to the training container.

Question 12

A company is using Amazon SageMaker to run a hyperparameter tuning job for an XGBoost model. The tuning job must efficiently search a large hyperparameter space while minimizing training time and cost. Which two strategies should the company use to achieve these goals? (Choose two.)

Accepted Answer

Enable early stopping in the hyperparameter tuning job to stop poorly performing training jobs before they complete.. Bayesian optimization is sample-efficient and finds good hyperparameters with fewer trials, reducing cost and time. Early stopping terminates underperforming trials early, further saving resources. Together, they efficiently search a large space while minimizing training time. Grid search and excessive job counts increase cost, while random search with too few jobs may not find optimal hyperparameters.

Answer

Set the maximum number of training jobs to a very high value to ensure thorough exploration.

Answer

Use grid search to exhaustively evaluate all possible hyperparameter combinations.

Answer

Use random search with a small number of training jobs to quickly sample the hyperparameter space.

Question 13

A data engineering team is building a pipeline that uses AWS Glue to transform data from Amazon S3 and load it into Amazon Redshift. The team wants to ensure that the Glue job can read from S3 and write to Redshift securely. They have created an IAM role for the Glue job with permissions to access S3 and Redshift. However, when they run the job, it fails with an authorization error when attempting to write to Redshift. What is the most likely cause of this failure?

Accepted Answer

The IAM role lacks the redshift:GetClusterCredentials permission, and the Redshift database user does not have the necessary privileges to write to the target table.. When AWS Glue writes to Amazon Redshift, it uses the IAM role attached to the job to call redshift:GetClusterCredentials and obtain temporary database credentials. If that permission is missing, the job cannot authenticate to Redshift. Furthermore, the Redshift database user must have appropriate privileges (e.g., INSERT) on the target table. Network or S3 staging issues would produce different errors. Therefore, the combination of missing IAM permission and insufficient database privileges is the most likely cause.

Answer

The IAM role attached to the Glue job does not have permissions to assume the Redshift service role.

Answer

The Glue job's IAM role does not have s3:PutObject permission on the Redshift cluster's staging directory.

Answer

The Redshift cluster is in a different VPC than the Glue job, and the security groups do not allow traffic.

Question 14

A data engineering team is building a feature store for a machine learning pipeline on AWS. Raw features arrive in Amazon S3 as Parquet files partitioned by event date, and the team wants to catalog them so that both Athena and Amazon SageMaker Feature Store can reference a consistent schema. The team also needs to keep the catalog updated automatically as new daily partitions arrive. Which two actions should the team take to meet these requirements? (Choose two.)

Accepted Answer

Define the table in the AWS Glue Data Catalog with the correct Parquet SerDe and partition keys so Athena and SageMaker Feature Store can reference a consistent schema.. Registering the Parquet data as a table in the AWS Glue Data Catalog with the proper SerDe and partition keys gives Athena and SageMaker Feature Store a shared schema definition. A scheduled Glue crawler, or partition projection, then discovers new date partitions automatically, keeping the catalog current as daily files land in S3.

Answer

Use AWS Glue DataBrew to profile the Parquet files and publish the profile as the table schema in the Data Catalog.

Answer

Store the Parquet files in Amazon Redshift Spectrum's external schema and let Redshift manage the partition metadata for Athena.

Answer

Enable S3 Object Lock on the bucket to guarantee that the Parquet schema cannot change once written.

Question 15

A team is training a SageMaker built-in Linear Learner model for a regression task predicting house prices. The training data contains a few extreme outliers in the target variable. The team wants the model to be robust to these outliers while still using the built-in algorithm. Which predictor_type and loss combination should they configure?

Accepted Answer

predictor_type='regressor' with loss='absolute_loss'. SageMaker Linear Learner supports multiple losses per predictor type. For regression with outlier-prone targets, absolute loss reduces the influence of large residuals because its gradient is bounded in magnitude, unlike squared loss whose gradient scales with the residual. Keeping predictor_type as regressor preserves the continuous output. Classification modes are structurally incapable of predicting prices, so the loss choice within regressor mode is the deciding factor.

Answer

predictor_type='multiclass_classifier' with loss='softmax'

Answer

predictor_type='binary_classifier' with loss='logistic'

Answer

predictor_type='regressor' with loss='squared_loss'

Question 16

A company collects application logs in Amazon CloudWatch Logs. A data scientist needs to train a model on the last 90 days of these logs, and the training job runs on an Amazon SageMaker notebook instance. The data scientist wants the simplest way to export the log data into Amazon S3 in a format that a SageMaker training job can read directly. Which action should the data scientist take?

Accepted Answer

Create an export task in CloudWatch Logs to write the log group to an S3 bucket, then point the training job at the exported S3 prefix.. CloudWatch Logs export tasks write a specified time range of a log group directly to an S3 bucket, producing objects that a SageMaker training job can consume without transformation. Because the data scientist needs historical logs from the past 90 days and wants minimal effort, the native export task is the only option listed that both backfills historical data and lands it in S3 in one operation.

Answer

Use Amazon Kinesis Data Firehose to stream the log group into Amazon S3 with a 90-day buffer.

Answer

Subscribe the log group to an AWS Lambda function that writes each log event to Amazon S3.

Answer

Query the log group with CloudWatch Logs Insights and copy the results into an Amazon S3 bucket manually.

Question 17

A data scientist is training a linear learner model on a dataset with 50 features. The model is overfitting, and the scientist wants to apply L1 regularization to encourage sparsity in the model weights. Which SageMaker built-in Linear Learner hyperparameter should be adjusted to control the strength of L1 regularization?

Accepted Answer

l1. The SageMaker built-in Linear Learner algorithm provides 'l1' and 'l2' hyperparameters for regularization. To encourage sparsity, the scientist should increase the 'l1' value, which adds an L1 penalty to the loss function and drives many weights to zero. L2 regularization ('l2') reduces overfitting but does not produce sparse models. 'wd' is not a valid hyperparameter, and 'learning_rate' affects optimization speed, not regularization.

Answer

learning_rate

Answer

l2

Answer

wd

Question 18

A data scientist is tuning a SageMaker XGBoost binary classifier and suspects the model is overfitting: training AUC is 0.99 while validation AUC plateaus near 0.82. The scientist wants to reduce the gap using hyperparameters exposed by the built-in algorithm. (Choose two.)

Accepted Answer

Increase the min_child_weight value.. Both min_child_weight and gamma constrain how splits are accepted, forcing each split to help a meaningful amount of data or achieve a real loss reduction. That reduces the tree's ability to memorize individual training rows, which is the usual source of a large train-validation gap. Capacity-increasing knobs such as depth and learning rate move in the opposite direction.

Answer

Set the objective to multi:softmax.

Answer

Increase the learning_rate and reduce the number of rounds.

Answer

Increase the max_depth value.

Question 19

A machine learning team stores training data in Amazon S3 and runs nightly SageMaker training jobs that read the same 2 TB dataset from many instances across multiple epochs. Training is I/O bound and throughput to S3 varies, extending job runtimes. The team wants to reduce data transfer time and provide consistent read performance. Which approach is MOST appropriate?

Accepted Answer

Attach an Amazon FSx for Lustre file system linked to the S3 bucket and use it as the training input channel.. FSx for Lustre is purpose-built for high-performance parallel access and integrates directly with S3, so a linked file system presents the dataset as a fast, shared POSIX namespace. SageMaker training jobs can mount it as an input channel, giving many instances consistent throughput across repeated epochs. EFS, EBS copies, and Transfer Acceleration do not deliver the same combination of shared access and sustained throughput.

Answer

Create an Amazon EFS file system and mount it to the training instances as the input data source.

Answer

Enable S3 Transfer Acceleration on the bucket and continue reading directly from S3.

Answer

Copy the dataset to the instance's Amazon EBS volume before each epoch using a lifecycle script.

Question 20

A data scientist suspects that the 'signup_date' timestamp column in a customer dataset contains parsing problems, because several rows fail downstream feature extraction. The scientist wants to inspect the detected data types, missing-value counts, and value distributions for that column without writing transformation code. Which capability should be used?

Accepted Answer

Configure a SageMaker Data Wrangler data quality and insights report on the dataset. The requirement is code-free profiling of a suspicious column: inferred type, missing values, and distribution. The Data Wrangler data quality and insights report provides precisely that profiling output as a managed job. Ground Truth is for labeling, Model Monitor is for deployed-endpoint drift detection, and Debugger observes training tensors, so none of them profile an offline column for parsing problems.

Answer

Run a SageMaker Debugger rule on the dataset to detect the timestamp anomalies

Answer

Create a SageMaker Ground Truth labeling job with a date-annotation task type

Answer

Enable SageMaker Model Monitor on the dataset and read the data quality baseline violations