MLA-C01 Practice Test 39 — 15 Questions

Question 1

An organization uses SageMaker Studio and needs to restrict Studio's internet access while allowing users to install custom packages from a private PyPI mirror hosted in a VPC. Which networking configuration should they use?

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

Accepted Answer

Disable internet access for Studio and configure VPC-only mode, then use a VPC endpoint to the private PyPI mirror. To restrict Studio's internet access while allowing access to a private PyPI mirror in a VPC, you must disable internet access for Studio and configure VPC-only mode, then create a VPC endpoint (e.g., an interface endpoint or a private link) to the mirror. This keeps all traffic within the VPC and avoids public internet exposure. The VPC endpoint provides private connectivity to the mirror without a NAT gateway or internet gateway.

Answer

Use a NAT gateway to allow outbound traffic to the private PyPI mirror

Answer

Disable internet access for Studio and rely on SageMaker's default VPC configuration

Answer

Enable internet access for Studio and use a VPC endpoint to the private PyPI mirror

Question 2

A data scientist wants to fine-tune a large language model for a question-answering task. They want to reduce memory usage during training by using a low-rank approximation of the weight updates. Which technique should they use?

Accepted Answer

LoRA. LoRA (Low-Rank Adaptation) adds low-rank matrices to model weights, significantly reducing memory footprint while achieving competitive performance. QLoRA adds quantization for further reduction.

Answer

Full fine-tuning

Answer

Instruction tuning

Answer

RLHF

Question 3

A data scientist wants to normalize a feature to have a range between 0 and 1 for a neural network. Which scaling technique should be applied?

Accepted Answer

MinMaxScaler. MinMaxScaler scales each feature to a given range, typically [0,1], by subtracting the minimum and dividing by the range. This makes it the correct choice for normalizing data to [0,1]. RobustScaler uses median and IQR, StandardScaler uses mean and standard deviation (resulting in zero mean, unit variance), and MaxAbsScaler scales to [-1,1] by dividing by the maximum absolute value. Therefore, only MinMaxScaler guarantees a [0,1] range.

Answer

RobustScaler

Answer

StandardScaler

Answer

MaxAbsScaler

Question 4

Refer to the exhibit. An IAM policy is attached to a user to allow invoking a SageMaker endpoint. A developer tries to call the endpoint from a laptop with IP 203.0.113.5 and receives an access denied error. What is the most likely reason?

Accepted Answer

The condition restricts the IP address to the 10.0.0.0/8 range.. The policy includes a condition that restricts the source IP address to the 10.0.0.0/8 private range. The developer's laptop has a public IP of 203.0.113.5, which does not fall within that range, so the condition fails and access is denied. This is the most likely reason for the error because the condition explicitly blocks requests from outside the specified private network.

Answer

The resource ARN is incorrect.

Answer

The user does not have permission to assume the SageMaker role.

Answer

The policy does not include access to the API action.

Question 5

A machine learning team needs to automatically retrain a model when concept drift is detected in the deployed endpoint's predictions. Which TWO steps should they take? (Choose TWO.)

Accepted Answer

Create a CloudWatch alarm on a model quality metric (e.g., accuracy) and trigger a Lambda function to start a retraining job. Option B is correct because a CloudWatch alarm on a model quality metric such as accuracy can detect when the deployed model's performance degrades, and its alarm action can invoke a Lambda function that programmatically starts a SageMaker retraining job, giving the automatic, event-driven retraining the team requires. Option C is correct because SageMaker Model Monitor's Model Quality Monitor computes prediction quality metrics (accuracy, precision, recall, etc.) by comparing captured endpoint predictions against ground truth labels, which is exactly the mechanism needed to detect concept drift in the model's predictions. Option A is not appropriate because a fixed EventBridge schedule retrains regardless of whether drift actually occurred, which is time-based rather than drift-triggered. Option D is incorrect because the Data Quality Monitor detects input/data drift (changes in the feature distribution), not concept drift in prediction quality. Option E is incorrect because SageMaker Clarify is used for bias detection and explainability, not for monitoring concept drift in model predictions.

Answer

Schedule retraining with Amazon EventBridge on a fixed schedule

Answer

Configure SageMaker Model Monitor - Data Quality Monitor to detect input drift

Answer

Use SageMaker Clarify to monitor bias drift

Question 6

A machine learning team needs to deploy a PyTorch model that has been compiled with SageMaker Neo to improve inference performance on edge devices. Which TWO statements about SageMaker Neo are correct? (Select TWO.)

Accepted Answer

Neo reduces model inference latency through optimization techniques. Option A is correct because SageMaker Neo applies compiler-level optimizations such as operator fusion, constant folding, and quantization-aware graph rewrites that reduce inference latency and model size on the target hardware. Option C is correct because Neo is a compilation service that takes a trained model plus a target hardware specification (for example, an Intel x86 CPU with a specific instruction set, an ARM Cortex-A SoC, or an NVIDIA Jetson GPU) and produces a hardware-optimized executable, so the compiled artifact is tied to that target. Option B is incorrect because Neo can compile models trained anywhere, including on-premises or in other clouds, as long as the model is provided in a supported framework format such as PyTorch, TensorFlow, MXNet, or ONNX. Option D is incorrect because Neo supports custom models and frameworks, not just SageMaker built-in algorithms. Option E is incorrect because endpoint auto scaling is handled by Application Auto Scaling and SageMaker endpoint scaling policies, not by Neo, which only handles model compilation.

Answer

Neo requires the model to be trained on SageMaker

Answer

Neo can only compile models trained with SageMaker built-in algorithms

Answer

Neo automatically scales SageMaker endpoints based on demand

Question 7

A machine learning engineer is preparing a training job on SageMaker with a custom Docker container. Which TWO actions are required to use the container with SageMaker? (Choose TWO.)

Accepted Answer

Push the container image to Amazon ECR. Option A is correct because SageMaker can only pull custom training container images from Amazon ECR, so the image must be built and pushed to an ECR repository that the SageMaker execution role can access. Option B is correct because the SageMaker Estimator must be configured with the image_uri parameter set to the ECR image URI (for example, .dkr.ecr..amazonaws.com/:) so SageMaker knows which container to run for training. Option C is incorrect because container images cannot be stored or executed from Amazon S3; S3 is used for training data and model artifacts, not Docker images. Option D is incorrect because SageMaker Debugger is an optional monitoring and profiling feature, not a requirement for using a custom container. Option E is incorrect because the SageMaker Model Registry is used to catalog trained models for governance and deployment, and is not needed to run a custom training container.

Answer

Upload the container image to Amazon S3

Answer

Enable SageMaker Debugger to monitor the custom container

Answer

Register the container in SageMaker Model Registry

Question 8

A data scientist is preparing a large dataset for training a machine learning model. The dataset contains missing values in several columns. Which approach is the MOST efficient for handling missing values in a large dataset using AWS services?

Accepted Answer

Use Amazon SageMaker Data Wrangler to impute missing values using built-in transforms.. Amazon SageMaker Data Wrangler provides a visual interface and built-in transforms for handling missing values efficiently at scale, without writing custom code. Glue ETL is more code-heavy, and imputation with pandas is not scalable for large datasets. Removing all rows with missing values is not always optimal and may not be efficient.

Answer

Use AWS Glue ETL to write a custom Python script that imputes missing values with the mean.

Answer

Use pandas in a SageMaker notebook to impute missing values with the median.

Answer

Remove all rows with missing values from the dataset.

Question 9

A data scientist is using Amazon SageMaker Debugger to monitor training metrics. They want to stop training automatically if the model is overfitting. Which action should they take?

Accepted Answer

Configure a custom rule that triggers a STOP training action when validation loss stops decreasing. SageMaker Debugger allows you to define custom rules that can invoke a STOP training action when a specified condition is met, such as validation loss ceasing to decrease. This enables automatic termination of a training job to prevent overfitting, as the model is no longer improving on unseen data.

Answer

Define a Debugger rule that monitors the loss plateau

Answer

Create a SageMaker Training Compiler

Answer

Use a built-in rule that checks for vanishing gradients

Question 10

A company wants to deploy 50 small models (each ~100 MB) for real-time inference. They need to minimize hosting costs while maintaining low latency. Which SageMaker hosting option is most cost-effective?

Accepted Answer

SageMaker Multi-Model Endpoint (MME). SageMaker Multi-Model Endpoint (MME) hosts many models on a single endpoint and dynamically loads them into memory/disk on invocation, sharing the underlying instance. For 50 small models (~100 MB each), this dramatically reduces hosting cost versus one endpoint per model while still providing real-time, low-latency inference. Serverless Inference is pay-per-invoke but has cold starts and memory limits that can hurt latency for frequent calls.

Answer

SageMaker Serverless Inference

Answer

SageMaker Asynchronous Inference

Answer

SageMaker real-time endpoint with one instance per model

Question 11

A machine learning engineer trains a binary classifier and obtains an accuracy of 95% on the test set. The dataset is imbalanced with 95% positive class. What is the most important metric to evaluate the model's performance?

Accepted Answer

F1 score. With a 95% positive class imbalance, a model that always predicts the majority class achieves 95% accuracy, making accuracy a misleading metric. The F1 score (option B) is the harmonic mean of precision and recall, providing a balanced evaluation of the model's ability to correctly identify the minority class while penalizing false positives and false negatives. This makes it the most important metric for imbalanced binary classification.

Answer

R-squared

Answer

Accuracy

Answer

RMSE

Question 12

Which TWO data storage options are commonly used by Amazon SageMaker Feature Store for offline and online storage?

Accepted Answer

Amazon S3. Amazon S3 (option D) is correct because SageMaker Feature Store uses it as the underlying offline store, where feature data is written in Parquet format for historical/batch retrieval and training. Amazon DynamoDB (option E) is correct because it serves as the online store, providing low-latency, high-throughput reads of the latest feature values for real-time inference. Amazon Redshift (A), Amazon RDS (B), and Amazon ElastiCache (C) are not the managed storage backends used by Feature Store for its online and offline stores, even though they are valid AWS data services in other contexts.

Answer

Amazon Redshift

Answer

Amazon RDS

Answer

Amazon ElastiCache

Question 13

An ML team wants to use Amazon SageMaker Ground Truth to create a labeled dataset for a multi-class image classification task. They have a large set of unlabeled images and want to minimize labeling costs while maintaining high accuracy. Which Ground Truth feature should they enable?

Accepted Answer

Active learning. Active learning in SageMaker Ground Truth automatically selects the most informative unlabeled images for human labeling, reducing the total number of labels needed while maintaining model accuracy. By iteratively training a model on a small labeled subset and then using that model to identify uncertain predictions, the system focuses labeling effort on the data that will most improve the model, directly minimizing labeling costs.

Answer

Annotation consolidation

Answer

Data labeling workforce management

Answer

Consolidated labeling

Question 14

A data scientist is using SageMaker Autopilot to automatically build a model. Which TWO aspects does Autopilot handle? (Choose TWO.)

Accepted Answer

Feature engineering. SageMaker Autopilot automatically performs feature engineering (option C), including data preprocessing, transformations, and generating candidate feature pipelines as part of exploring the dataset and building the best model. It also automatically performs hyperparameter tuning (option E), running many training jobs with different algorithm and hyperparameter combinations to optimize model performance. Autopilot does not handle data ingestion (option A), which is the responsibility of the user to provide data in S3 or another source before starting the job. It also does not perform model deployment (option B), which must be done separately via SageMaker endpoints or batch transform after Autopilot produces the model. Finally, it does not perform data labeling (option D), since Autopilot requires an already-labeled target column for supervised learning.

Answer

Data ingestion

Answer

Model deployment

Answer

Data labeling

Question 15

Which TWO options are recommended best practices for monitoring model performance in production on SageMaker? (Choose 2.)

Accepted Answer

Use SageMaker Clarify for bias monitoring and feature importance drift.. Option B is correct because SageMaker Clarify is the purpose-built service for detecting bias in models and monitoring feature importance drift, which are essential aspects of production model performance monitoring beyond raw accuracy. Option C is correct because SageMaker Model Monitor is designed to automatically capture data drift, model quality metrics, and other distributional changes by comparing live endpoint traffic against a baseline, making it a recommended best practice for ongoing production monitoring. Option A is not recommended because retraining daily without evaluation risks deploying degraded or biased models and is not a monitoring practice. Option D is insufficient as a best practice because the Invocations metric only counts request volume and reveals nothing about model quality, drift, or bias. Option E is not recommended because manual weekly comparisons are ad hoc, error-prone, and not scalable compared to automated monitoring services.

Answer

Retrain the model daily based on recent data without evaluation.

Answer

Set up a CloudWatch alarm on the endpoint's Invocations metric.

Answer

Manually compare prediction distributions weekly.