AWS Certified Machine Learning Engineer – Associate (MLA-C01)
Exam Notes & Practice Tests

A technology company has recently migrated its artificial intelligence (AI) workloads to AWS Cloud and is looking for an optimized hardware solution to improve performance and cost-efficiency. The company primarily focuses on deep learning training and inference tasks for various AI applications, including generative AI models. Which of the following AWS services should the company use for optimizing hardware selection for AI workloads?

You are a machine learning engineer at a biotech company training a deep learning model for genomic sequence analysis. The dataset consists of billions of DNA sequences, requiring high computational power for training. After deployment, the model must provide real-time inference for clinical decision-making with minimal latency. Which AWS instance configuration is MOST SUITABLE for both training and inference of this model?

A healthcare company needs to fine-tune a large language model (LLM) for summarizing medical reports. The company requires a no-code or low-code approach to minimize development effort while ensuring scalability and operational efficiency. The solution should be implemented using Amazon SageMaker. Which of the following services should be used to accomplish this task? (Select four)

A healthcare analytics company is developing a machine learning model to predict patient readmission risks. The dataset includes structured tabular data, such as admission history and medical charges, alongside personally identifiable information (PII) like social security numbers and phone numbers. The company must ensure that sensitive information is masked before sharing the dataset for model training. Which AWS service should be used to meet these requirements?

A financial services company is improving its fraud detection system and marketing segmentation. The company has a large dataset containing features like transaction volume, frequency, and location. The ML team must use Amazon SageMaker built-in algorithms to complete the following tasks: Which combination of SageMaker built-in algorithms should the team use?

• Reduce dimensionality to enhance visualization and model performance.
• Identify customer clusters for targeted marketing campaigns.
• Detect anomalous transactions that may indicate fraud.

A data scientist is preparing datasets for an ML model and needs to handle missing values, duplicate records, and outliers. The data scientist must also merge multiple datasets into a single data frame for further preprocessing. Which is the most efficient AWS solution for this task?

An e-commerce company runs multiple ML models for recommendation systems, demand forecasting, and customer segmentation. To ensure performance optimization, a centralized dashboard is needed to monitor model latency, accuracy, and resource utilization across different environments. Which AWS service is BEST suited for creating a user-friendly monitoring dashboard?

A financial institution is running an ML-powered fraud detection model on Amazon SageMaker. Users have reported increased prediction latency and occasional timeout errors. To diagnose and resolve the issue, which approach should be used for monitoring and alerting?

A research lab is using Amazon SageMaker to train an ML model with millions of small files stored in Amazon S3. Training is taking longer than expected due to inefficient data access. Which immediate solution would enhance training performance?

A hospital is using computer vision to monitor access to restricted areas. Cameras capture images of individuals entering the facility, and the hospital needs to automatically verify whether an individual is an authorized staff member. The hospital has limited images of authorized personnel, making it impractical to train a custom model from scratch. Which AWS services should be used to implement the solution? (Select two)

A retail company is developing a deep learning model for personalized product recommendations. The model requires substantial training on customer behavior data, and the company wants to ensure optimal training performance. How does model training work in deep learning?

A tech company is developing an AI-driven chatbot using generative models. The company needs to distinguish between discriminative and generative models to determine the best approach for its AI system. What is the key difference between these two model types?

A machine learning engineer is developing a regression model to estimate vehicle resale prices based on features such as mileage, model year, and condition. However, the model exhibits overfitting due to high variance in certain features. The engineer is considering applying L1 or L2 regularization to mitigate this issue. Which regularization technique should be used, and why?

A software company is working on an ML-powered analytics dashboard that requires real-time access to operational data stored in Amazon DynamoDB. The company wants to integrate this data with a SageMaker notebook for model development. What is the MOST EFFICIENT way to access DynamoDB data from Amazon SageMaker?

A logistics company has developed a machine learning model to estimate package delivery times based on real-time traffic data. The company needs to deploy the model in production while meeting the following requirements: Which deployment method best satisfies these requirements?

• Low latency for real-time predictions
• Scalability to handle peak-hour surges
• High availability to ensure continuous service
• Dynamic scaling to optimize costs

A customer support chatbot powered by a large language model (LLM) in Amazon Bedrock is inconsistently responding to frequently asked questions about company policies. The business wants to ensure the chatbot provides more consistent responses. What is the BEST approach to achieve this?

A pharmaceutical company is using machine learning to analyze drug effectiveness. The dataset consists of 6 TB of structured and unstructured medical records stored in Amazon S3. The preprocessing phase includes intensive feature engineering and data transformations requiring distributed computing. The company wants to automate the full ML workflow. Which AWS service provides the best solution?

An organization is looking for a distributed computing solution to preprocess large-scale datasets for machine learning. The data is spread across multiple sources, requiring parallel processing capabilities. Which AWS service provides the best solution?

A cybersecurity company uses an API to retrieve threat intelligence reports from third-party sources. To enhance security, the company must implement automatic API token rotation every 90 days. What is the most effective way to achieve this?

A retail company has trained a sales forecasting ML model using Amazon SageMaker. The training data was preprocessed using z-score normalization in AWS Glue DataBrew. Now, the company must ensure that real-time sales data is preprocessed identically before making predictions. Which approach is the most effective?

A research organization is using Amazon SageMaker to train a deep learning model for detecting environmental changes in satellite images. The dataset, comprising 10 TB of high-resolution images, is stored in Amazon FSx for Lustre, which is linked to an Amazon S3 bucket. The machine learning team wants to ensure the most efficient access to training data while minimizing storage latency. Which is the BEST way to access the dataset during training?

A financial services company is working on an ML project that involves analyzing high-frequency transaction data. The data team wants to store and transform features efficiently before using them in model training within Amazon SageMaker. Which approach ensures optimal feature storage and transformation for machine learning models?

A data science team is using Amazon SageMaker Studio to develop ML models for market trend predictions. The company needs to set up an automated alert system to notify administrators when the SageMaker compute costs exceed a specified threshold. Which solution will provide the BEST cost monitoring and alerting mechanism?

Which statement correctly differentiates feature engineering for structured and unstructured data?

A fraud detection team at a payment processing company has built an ML model in Amazon SageMaker. The model generates a confidence score for each transaction. The company wants to receive email notifications when the fraud confidence score falls below 75%. Which of the following solutions should the ML team implement? (Select two.)

An AI research lab is training a deep learning model on Amazon SageMaker to classify satellite images. Due to large-scale computations, the energy consumption has increased significantly. The company wants to implement sustainable practices to reduce energy costs while maintaining model accuracy. Which strategies will help optimize training efficiency and reduce energy consumption? (Select two.)

In machine learning, what is the bias-variance tradeoff?

A retail company is adopting AI/ML for personalized recommendations and is comparing Amazon Bedrock with Amazon SageMaker JumpStart. What is the key difference between these two AWS services?

How do K-Means and K-Nearest Neighbors (KNN) differ in machine learning?

An e-commerce company uses machine learning to generate product descriptions for images uploaded by sellers. Since each uploaded image can be as large as 50 MB, the company needs a scalable and cost-efficient inference solution that adapts to fluctuating demand. Which approach is the most efficient?

A cloud-based fintech company is designing a microservices architecture for its fraud detection and financial analytics platform. The company wants to define its infrastructure as code (IaC) while ensuring scalability and automation. Additionally, it must deploy containerized ML models that dynamically scale to handle varying workloads. Which combination of IaC and container orchestration service is the MOST SUITABLE for this scenario?

A telecommunications company is developing an ML model to predict customer churn based on structured customer data, including billing history, service usage, and support interactions. The dataset is stored in Amazon S3 in CSV format. The ML engineer must choose an Amazon SageMaker built-in algorithm that supports tabular data and is designed for binary classification. Which algorithm should the ML engineer use?

An e-commerce platform streams real-time user interactions using Amazon Kinesis Data Streams to analyze customer behavior and product recommendations. The company needs to aggregate the data at one-minute intervals to compute features like session duration and average page views per session with minimal latency. Which combination of solutions BEST meets these requirements? (Select two.)

You are using SageMaker Data Wrangler to prepare a dataset before training a machine learning model. The dataset has missing values in numeric columns, and you need to replace them with the column mean to ensure smooth model training. Which transformation operation should you apply?

A streaming media company logs daily video views in Amazon S3 and uses Amazon Athena to generate weekly reports on viewing trends. The company wants to retain the data for 90 days before moving it to low-cost storage. The solution must optimize Athena query performance while maintaining cost efficiency. What is the BEST approach?

A genomics research organization has encrypted genomic datasets stored in an Amazon S3 bucket using AWS KMS (SSE-KMS). A data scientist needs to analyze these datasets in an Amazon SageMaker notebook while ensuring secure decryption of the files. Which actions ensure secure data access? (Select two.)

A marketing analyst has built an ML model using an external framework and stored the model artifacts in an Amazon S3 bucket. The analyst wants to share the model with a colleague using Amazon SageMaker Canvas. What conditions must be met? (Select two.)

An ML engineer has deployed a fraud detection model using Amazon SageMaker and needs to monitor model drift and track accuracy in real-time. Which AWS service is best suited for this task?

A machine learning engineer manages a real-time fraud detection model deployed on Amazon SageMaker. The model must handle high-volume transactions with low latency, but latency spikes have increased during peak hours. What is the BEST way to resolve these latency and scaling issues?

A logistics company needs to deploy an ML model for daily demand forecasting. The workload occurs in a 90-minute window each day, requiring low latency responses during high concurrent invocations. The company wants AWS to manage scaling with minimal infrastructure maintenance. Which cost-effective inference option is best?

A data scientist at a fintech company is developing an ML model to detect suspicious transactions. The dataset is large and stored in Amazon S3, requiring batch processing for inference. The company also wants to continuously monitor data quality and be alerted in case of significant changes. Which solution should the data scientist implement?

A healthcare AI company is training a deep learning model for medical image analysis using distributed training on multiple GPU instances. The training process is slower than expected due to network communication overhead. Which actions should the ML engineer take to optimize network performance? (Select two.)

A biotech research company has been training ML models on-premises using PyTorch and proprietary medical datasets. The company wants to migrate its ML workflows to AWS while making minimal changes to its custom PyTorch scripts. Which AWS service is the best fit for migrating the existing workflows?

A news media company wants to implement semantic search on a large document archive stored in Amazon S3. Journalists should be able to search for articles using natural language queries. Which AWS service provides the best solution for semantic search?

A travel agency is using Amazon Q Business to answer customer queries about vacation packages. The company must ensure that outdated destinations are not included in API responses. How can the company filter outdated travel destinations from API responses?

A telecom provider is building an ML model to predict customer churn using a dataset with hundreds of features. The data science team wants to reduce model complexity by focusing only on the most important features. Which technique should the team use to select the most relevant features?

A healthcare startup is developing an ML model to detect rare diseases. False negatives (failing to detect an actual case) could have severe consequences, while false positives (incorrectly classifying a patient as diseased) are less harmful. Which metric should the company prioritize?

A medical research firm is using Amazon SageMaker XGBoost to classify patients as high-risk or low-risk for a disease. The model performs well on training data but poorly on unseen data, likely due to overfitting. What should the ML engineer do to improve generalization?

A streaming service wants to test a new recommendation model before full deployment. 20% of user traffic should be routed to the new model, while the rest continues using the existing model. Which SageMaker deployment strategy meets this requirement?

A media company uses AWS Lake Formation to manage its data lake containing structured and unstructured data such as video metadata, user logs, and clickstream data. Analysts are assigned to specific content categories, such as sports, movies, or documentaries. The company wants to ensure fine-grained access control while simplifying administration. Which solution BEST meets these requirements?

A financial services company processes sensitive customer data, including Social Security numbers and financial transactions, stored in Amazon S3. The company must automatically identify and remove sensitive data before processing it with an on-premises ML model. The solution should require minimal operational overhead. Which AWS service provides the best solution for this use case?

A retail company is building a recommendation engine based on historical user behavior, including product views, purchases, and ratings. The ML engineer must choose a built-in SageMaker algorithm for training a personalized recommendation model with minimal development effort. Which built-in SageMaker algorithm is the most suitable?

A healthcare company uses multiple scikit-learn models to predict disease risk factors and needs to deploy them in Amazon SageMaker for low-latency predictions while optimizing costs. Which solution best meets these requirements? (Select two.)

A financial services company retrains its fraud detection ML model weekly using Amazon SageMaker Pipelines. The pipeline includes data preprocessing, model training, evaluation, and model registration. The preprocessing step requires large-scale data transformations using Amazon EMR. How can the team integrate Amazon EMR with SageMaker Pipelines? (Select two.)

A retail company retrains its customer demand prediction model when new data is uploaded to Amazon S3. Currently, the SageMaker pipeline is manually triggered. The ML engineer must automate the process when new data is available. Which solution requires the least operational effort?

A healthcare AI company runs containerized ML applications on Amazon EC2, ECS, and Lambda. The company needs to identify underutilized resources and receive cost optimization recommendations. Which AWS service provides the most efficient solution?

A data scientist has deployed a sentiment analysis model in Amazon SageMaker and needs to explain its predictions to multiple stakeholders. Which AWS service provides model interpretability?

A SageMaker training job runs in a public subnet within an Amazon VPC. The company has detected malicious traffic from a specific IP address targeting the VPC. How should the company block traffic from this specific IP address while allowing legitimate traffic?

A banking institution stores sensitive customer data in Amazon Redshift. A data analyst requires access to customer behavior insights, but the company must restrict access to personally identifiable information (PII) without duplicating data. Which solution requires the least implementation effort?

A fintech company operates a real-time credit scoring model and needs to monitor availability, utilization, throughput, and fault tolerance to ensure high performance and reliability. Which combination of metrics and monitoring strategies is the most effective?

A media streaming company is building an ML recommendation model in Amazon Redshift ML in its primary AWS account. However, the training data is stored in Amazon S3 in a secondary AWS account. The company requires a secure solution to access this data without exposing it to the public internet. Which approach should the company take?

A retail business relies on an ML model to forecast daily sales and adjust inventory levels accordingly. The model runs once every evening, requires 3.5 MB of input data, and completes within 60 seconds. The company needs an optimized inference deployment on Amazon SageMaker that balances cost and efficiency. Which SageMaker inference option is the most suitable?

A financial institution has developed an ML model for credit score predictions and must comply with regulatory guidelines on transparency, fairness, and security. The ML team needs an AWS service that provides: Which AWS service is best suited for ensuring ML model governance?

• Model explainability
• Bias detection
• Governance tracking

A customer service organization wants to predict whether future clients will require long-term support based on historical data. The company needs to select the most appropriate machine learning approach to model this problem. Which ML approach should the company use?

You are developing a loan approval model for a bank. The model must ensure fairness and avoid discrimination against age, gender, or ethnicity. The evaluation must assess both model performance and bias detection. Which combination of metrics and bias detection methods is most appropriate?