MLS-C01 Practice Exam 57 — 20 Questions

Question 1

A team is exploring a dataset with missing values in multiple columns. They want to decide whether to drop rows or impute values. Which approach is most appropriate for exploratory data analysis?

Accepted Answer

Analyze the missing data pattern using visualizations and summary statistics. During EDA, the goal is to understand the nature and pattern of missingness before deciding on a treatment. Visualizing missingness (e.g., missingno matrix, heatmap) and computing summary statistics (per-column null counts, correlation of missingness with other features) reveals whether data is MCAR, MAR, or MNAR — which dictates whether dropping or imputing is appropriate. Jumping straight to imputation or row deletion without this diagnostic step risks introducing bias or discarding valuable signal.

Answer

Impute missing values with the mean of each column

Answer

Drop all rows with missing values to ensure data quality

Answer

Use Amazon SageMaker Data Wrangler to automatically impute missing values

Question 2

Refer to the exhibit. A SageMaker training job is launched with the CLI command shown. The job fails with an error 'S3 data distribution type not supported for File mode'. What is the most likely fix?

Accepted Answer

Change TrainingInputMode to Pipe. The error 'S3 data distribution type not supported for File mode' occurs because SageMaker's File mode downloads the entire dataset to the training instance's local storage before training begins, which does not support the ShardedByS3Key data distribution type. Changing TrainingInputMode to Pipe mode resolves this, as Pipe mode streams data directly from S3 to the algorithm, allowing each instance to receive a unique shard of the data without requiring local storage.

Answer

Change InstanceCount to 2

Answer

Increase VolumeSizeInGB to 100

Answer

Increase MaxRuntimeInSeconds to 7200

Question 3

A data engineering team needs to ingest streaming data from thousands of IoT devices into a data lake on Amazon S3 for near-real-time analytics. The data must be partitioned by device ID and timestamp, and the team must minimize data loss during ingestion failures. Which solution is MOST appropriate?

Accepted Answer

Use Amazon Kinesis Data Firehose to write directly to S3 with dynamic partitioning.. Amazon Kinesis Data Firehose with dynamic partitioning is the most appropriate solution because it natively supports partitioning incoming data by device ID and timestamp before writing to S3, and it provides built-in data buffering and retry logic to minimize data loss during ingestion failures. Unlike a Lambda-based approach, Firehose handles large-scale streaming ingestion without requiring custom code for partitioning or error handling, making it ideal for near-real-time analytics on IoT data.

Answer

Use Amazon Kinesis Data Streams with a Lambda function that writes to S3.

Answer

Use Amazon S3 Transfer Acceleration with direct uploads from devices.

Answer

Use AWS Lambda to receive data via API Gateway and write to S3.

Question 4

A machine learning engineer is exploring a dataset with 50 features. Some features are highly correlated. Which technique should the engineer use to reduce dimensionality while preserving variance?

Accepted Answer

Principal Component Analysis (PCA). Principal Component Analysis (PCA) is a dimensionality reduction technique that transforms the original features into a new set of uncorrelated components (principal components) ordered by the amount of variance they explain. By selecting the top components, the engineer can reduce dimensionality while preserving most of the variance in the data, which is exactly the requirement.

Answer

Factor Analysis

Answer

t-Distributed Stochastic Neighbor Embedding (t-SNE)

Answer

Linear Discriminant Analysis (LDA)

Question 5

A data scientist is training a neural network on Amazon SageMaker and wants to automatically stop training if the validation loss does not improve for 5 consecutive epochs. Which feature should they use?

Accepted Answer

Implement early stopping in the training script. Early stopping is a training technique implemented within the training script itself, where the algorithm monitors validation loss and halts training if no improvement is seen for a specified number of epochs (patience). In SageMaker, this is not a built-in service feature but must be coded by the data scientist using frameworks like TensorFlow, PyTorch, or Keras callbacks. Therefore, Option A is correct because it directly addresses the requirement.

Answer

SageMaker Debugger

Answer

SageMaker Checkpointing

Answer

SageMaker Hyperparameter Tuning

Question 6

A data scientist is training a random forest model. During hyperparameter tuning, which parameter is MOST effective at reducing overfitting?

Accepted Answer

Decrease the maximum depth of each tree. Decreasing the maximum depth of each tree limits the complexity of individual trees, preventing them from memorizing noise and outliers in the training data. This directly reduces overfitting by enforcing simpler decision boundaries, which is a core regularization technique for ensemble methods like Random Forest.

Answer

Increase the number of trees

Answer

Increase the number of features considered per split

Answer

Increase the maximum depth of each tree

Question 7

A company is using Amazon Kinesis Data Firehose to load streaming data into an S3 bucket. The data schema evolves over time, with new columns added. The data must be queryable using Amazon Athena. What is the BEST way to handle schema changes?

Accepted Answer

Use AWS Glue Crawlers to automatically detect schema changes and update the table metadata. AWS Glue Crawlers scan the S3 data, infer the schema, and automatically update the AWS Glue Data Catalog, which Athena uses as its metastore. When Firehose writes new columns into S3, the next crawler run detects the added fields and updates the table metadata, so Athena queries immediately reflect the new schema without manual intervention. This is the standard serverless pattern for schema evolution in an Athena-on-S3 pipeline.

Answer

Manually update the Athena table definition each time a new column is added

Answer

Configure Firehose to convert the data to Apache JSON format

Answer

Recreate the Athena table daily to pick up new columns

Question 8

You are a data engineer at a fintech company. The company processes real-time stock market data from multiple exchanges. The data is ingested via Amazon Kinesis Data Streams with 50 shards. Each record is about 1 KB, and the ingestion rate is 5,000 records per second. The data is consumed by a Java application running on Amazon ECS that performs real-time analytics and stores results in Amazon DynamoDB. Recently, the application has been experiencing high latency, and some records are stuck in the shards for minutes before being consumed. The CloudWatch metrics show that the application's CPU utilization is low, but the iterator age is increasing. The application uses the Kinesis Client Library (KCL) with a single worker. What is the most likely cause and how should it be fixed?

Accepted Answer

Scale the number of KCL workers to match the number of shards (e.g., 50 workers) to process shards in parallel.. The Kinesis Client Library (KCL) assigns one shard per worker thread/process; with a single worker and 50 shards, only one shard is processed at a time, causing iterator age to climb while CPU stays idle. Scaling to 50 KCL workers (one per shard) allows parallel consumption across all shards, eliminating the backlog. This matches the symptom of low CPU with rising iterator age — a classic under-provisioned consumer signature.

Answer

Increase the number of shards to 200 to provide more throughput.

Answer

Increase the CPU capacity of the ECS task by moving to a larger instance type.

Answer

Move the destination from DynamoDB to Amazon RDS to reduce write latency.

Question 9

A data scientist is working with a dataset containing geospatial coordinates (latitude and longitude) of customer locations. The scientist wants to engineer features such as distance to the nearest store, and cluster customers into regions. Which AWS service is best suited for performing geospatial analysis and clustering during exploratory data analysis?

Accepted Answer

Amazon SageMaker with custom Python scripts using scikit-learn and Geopy. Amazon SageMaker with custom Python scripts using scikit-learn and Geopy is the best fit because it gives the data scientist full programmatic control to compute distances (Geopy) and run clustering algorithms like K-Means or DBSCAN (scikit-learn) during exploratory data analysis. SageMaker provides the managed notebook and compute environment, while the libraries handle the geospatial math and clustering. This is the most flexible option for custom feature engineering.

Answer

Amazon Athena with PostGIS extensions

Answer

AWS Glue with geospatial transforms

Answer

Amazon Location Service

Question 10

A company uses Amazon Redshift for data warehousing. The data engineering team needs to load data from multiple S3 buckets into Redshift daily. Each bucket contains files in different formats (CSV, JSON, Parquet). Which AWS service is BEST suited to automate this ingestion process?

Accepted Answer

AWS Glue. AWS Glue is a fully managed ETL service that natively supports crawling S3 data sources, automatically detecting schemas for CSV, JSON, and Parquet formats, and generating transformation jobs that load data into Redshift. It integrates directly with the Glue Data Catalog and Redshift, allowing scheduled daily ingestion with minimal operational overhead. This makes it the best fit for automating multi-format, multi-bucket ingestion into Redshift.

Answer

Amazon EMR with Apache Spark

Answer

AWS Data Pipeline

Answer

AWS Database Migration Service (DMS)

Question 11

A data scientist is performing EDA on a dataset with both numerical and categorical features. Which technique is best for detecting multicollinearity among numerical features?

Accepted Answer

Correlation matrix with heatmap. A correlation matrix with heatmap computes pairwise Pearson (or Spearman) correlation coefficients between all numerical features, making it the standard EDA technique for spotting multicollinearity. High absolute values (typically |r| > 0.8 or 0.9) between two predictors signal redundancy that can destabilize linear models. The heatmap visualizes these coefficients so the analyst can quickly identify which features to drop or combine.

Answer

Chi-square test of independence

Answer

Box plots for each numerical feature

Answer

Pair plot

Question 12

Which THREE are common techniques for detecting outliers in a univariate dataset? (Select THREE.)

Accepted Answer

Z-score. The Z-score (C) is a standard univariate outlier detection technique that flags observations whose value deviates from the mean by more than a threshold number of standard deviations (commonly |z| > 3), directly operating on a single variable. The Interquartile range (IQR) method (D) is also a classic univariate approach: it computes Q1 and Q3, defines the bounds Q1 − 1.5·IQR and Q3 + 1.5·IQR, and marks points outside those fences as outliers. The Modified Z-score using median absolute deviation (E) is a robust univariate variant that replaces the mean and standard deviation with the median and MAD (typically using 0.6745·(x − median)/MAD), making it resistant to extreme values in a single variable. By contrast, Cook's distance (A) measures the influence of each observation on a regression model's fitted coefficients, so it is a multivariate/regression diagnostic rather than a univariate outlier test, and DBSCAN (B) is a density-based clustering algorithm that identifies outliers as noise points in a multi-dimensional feature space, not a univariate technique.

Answer

Cook's distance

Answer

DBSCAN clustering

Question 13

A data analyst needs to visualize the distribution of a numerical feature in a dataset. Which AWS service can be used to create a histogram directly from data stored in S3 without writing code?

Accepted Answer

Amazon QuickSight. Amazon QuickSight is a fully managed business intelligence service that allows users to create visualizations, including histograms, from data stored in Amazon S3 without writing code. It provides a drag-and-drop interface where you can connect to S3, import data, and build charts. QuickSight's SPICE engine enables fast, in-memory analysis, and it supports direct query to S3 via Athena or other sources. This makes it the ideal choice for a no-code visualization of a numerical feature's distribution.

Answer

Amazon Athena

Answer

Amazon SageMaker Studio

Answer

AWS Glue

Question 14

A fraud detection team needs features computed over a rolling 10-minute window from a high-volume stream of card transactions. The data currently lands in Amazon S3 via Amazon Kinesis Data Firehose, and a nightly AWS Glue job computes aggregates, so features are up to 24 hours stale and model accuracy suffers. The team wants features available to the online model within seconds of each event while reusing the same aggregation logic for both training and inference. Which solution meets these requirements?

Accepted Answer

Use Amazon Kinesis Data Analytics for Apache Flink to compute the windowed aggregates and write them to an Amazon SageMaker Feature Store online store.. A managed Apache Flink application can maintain event-time state and emit ten-minute rolling aggregates continuously from the stream, giving seconds-level freshness. Routing those features into the SageMaker Feature Store online store makes them retrievable with low latency at inference, while the feature group's offline store supplies point-in-time correct training data from the same definition. That single aggregation definition satisfies both freshness and reuse.

Answer

Increase the Firehose buffer interval to 60 seconds and have the model query Amazon Athena against the raw S3 objects.

Answer

Enable Kinesis Data Streams enhanced fan-out and have the model consume the raw events directly.

Answer

Run the existing AWS Glue job every 10 minutes with a smaller window and have the model read the output Parquet from Amazon S3.

Question 15

A financial services company is building a model to detect fraudulent credit card transactions. The dataset contains 1 million transactions, with only 0.1% labeled as fraud. The data scientist trains a logistic regression model on the raw dataset and obtains the following results on a held-out test set: accuracy = 99.8%, precision = 50%, recall = 60%, F1 = 0.545. The business requirement is to maximize recall while keeping precision above 80%. Which course of action should the data scientist take to improve the model?

Accepted Answer

Train the model with class weights inversely proportional to class frequencies. Assigning class weights inversely proportional to class frequencies penalizes misclassifications of the minority class (fraud) more heavily during training. This directly addresses the severe class imbalance (0.1% fraud) by forcing the logistic regression model to learn decision boundaries that improve recall, while the weight ratio can be tuned to maintain precision above 80%. Unlike naive resampling, this approach preserves the original data distribution and avoids information loss.

Answer

Use random undersampling of the majority class to balance the dataset

Answer

Collect more historical transaction data and retrain the model

Answer

Apply L2 regularization with a higher penalty to reduce overfitting

Question 16

A company uses Amazon SageMaker to train a model. The training job fails with an 'OutOfMemory' error. The training data is stored in S3 and the instance type is ml.m5.xlarge. What is the most efficient way to resolve this issue?

Accepted Answer

Use a larger instance type, such as ml.m5.2xlarge. The 'OutOfMemory' error indicates that the ml.m5.xlarge instance (4 vCPUs, 16 GiB memory) does not have enough RAM to hold the training data and model during processing. Upgrading to ml.m5.2xlarge (8 vCPUs, 32 GiB memory) directly increases available memory, resolving the issue without requiring code changes or architectural modifications. This is the most efficient solution because it requires no script alterations and leverages SageMaker's built-in instance scaling.

Answer

Enable managed spot training

Answer

Reduce the batch size in the training script

Answer

Increase the number of instances using distributed training

Question 17

A data engineer is using Amazon SageMaker Data Wrangler to perform exploratory data analysis on a large dataset stored in S3. The analysis reveals high cardinality in a categorical feature with over 1 million unique values. What is the best approach to handle this before training a model?

Accepted Answer

Use target encoding based on the mean of the target variable per category.. Target encoding (also called mean encoding) replaces each category with the average of the target variable for that category, which is a compact and effective way to handle high-cardinality categorical features without exploding dimensionality. It captures the relationship between the category and the target while keeping the feature space small. With over 1 million unique values, one-hot encoding would create an unusable number of columns, and label encoding would impose an arbitrary ordinal relationship.

Answer

Apply one-hot encoding.

Answer

Use label encoding to convert categories to integers.

Answer

Drop the high-cardinality feature.

Question 18

A company is training a deep learning model on Amazon SageMaker. The training job is failing with an out-of-memory error. Which SageMaker feature should the company use to resolve this issue without changing the instance type?

Accepted Answer

Use SageMaker distributed training with model parallelism. SageMaker's distributed training with model parallelism splits the model's layers across multiple GPUs, reducing the memory footprint per GPU. This allows the company to train a large model that exceeds a single GPU's memory without changing the instance type. Model parallelism is specifically designed to handle out-of-memory errors by distributing the model parameters, gradients, and optimizer states across devices.

Answer

Use SageMaker Savings Plans

Answer

Enable SageMaker Managed Spot Training

Answer

Use SageMaker Debugger to monitor memory usage

Answer

Enable SageMaker Profiler to profile memory

Question 19

A machine learning engineer is training a linear regression model on a dataset with 50 features. After training, the model achieves high accuracy on the training set but poor accuracy on the test set. Which technique should the engineer use to address this issue?

Accepted Answer

Apply L1 or L2 regularization. The model exhibits overfitting: high training accuracy but poor test accuracy. L1 (Lasso) or L2 (Ridge) regularization penalizes large coefficients, reducing model complexity and improving generalization. This directly addresses the variance problem without requiring more data or features.

Answer

Train a deeper neural network with more layers

Answer

Add more features through feature engineering

Answer

Increase the size of the training dataset

Question 20

A company uses Amazon SageMaker to host a model for real-time predictions. The model endpoint is experiencing high latency during peak hours. The data scientist wants to reduce latency without increasing cost. Which action should they take?

Accepted Answer

Enable auto-scaling for the endpoint based on latency metrics. Auto-scaling adds endpoint instances dynamically when latency (or invocations) crosses a threshold, spreading the load across more instances and reducing per-request latency without permanently provisioning larger or more expensive capacity. Because scaling is triggered only during peak hours, cost stays proportional to actual demand. This directly addresses the 'reduce latency without increasing cost' constraint.

Answer

Enable data capture for the endpoint to log requests

Answer

Switch to a larger instance type

Answer

Reduce the number of instances behind the endpoint