AWS Certified AI Practitioner Exam Notes and Practice Tests – Practice Test #1

You are a machine learning engineer at a biotech company that develops AI-driven diagnostic tools. Due to strict regulatory requirements, you must ensure that all models used for medical diagnoses are properly versioned, reproducible, and auditable. This includes tracking training data, hyperparameters, and model artifacts. Which strategy best ensures proper versioning, reproducibility, and auditability?

A startup is building an AI-based chatbot service that experiences unpredictable spikes in user queries. Some days have minimal traffic, while others experience a sudden surge. The company wants to minimize costs while maintaining the ability to scale dynamically in response to demand. Which of the following options provides the best balance between cost-efficiency and scalability?

A travel booking platform is deploying a machine learning model to personalize travel recommendations in real time. The system must provide low-latency responses and automatically scale during high-traffic periods such as holiday seasons. Which AWS service and configuration best meet these requirements?

A financial services company wants to automate its ML model retraining and deployment pipeline whenever new transaction data is uploaded to Amazon S3. The pipeline must ensure that retrained models are deployed seamlessly. Which three steps should be followed to configure this automation?

A pharmaceutical company needs a workflow orchestration solution to manage preclinical drug testing simulations using machine learning models. The system must support:
• Pipeline visualization as a directed acyclic graph (DAG)
• Experiment tracking for regulatory compliance
• Version control for model governance
Which AWS service best meets these requirements?

A social media company is using Amazon SageMaker Model Registry to store and manage ML models for different business units, including ad targeting, content recommendation, and fraud detection. The ML engineering team wants to organize models into business-specific categories while preserving model integrity. What is the best way to categorize models while maintaining their original groupings?

A gaming company has multiple ML models with different deployment needs:
• A real-time recommendation engine with low-latency requirements.
• A generative AI chatbot that must scale dynamically based on user load.
• A fraud detection model that should run within a serverless infrastructure.
Which deployment strategies are best suited for these models? (Select two)

A fintech startup is developing an ML model to detect fraudulent credit card transactions. The dataset is highly imbalanced, with only 2% of transactions being fraudulent. To improve detection accuracy, you consider using boosting techniques. Which boosting method is best suited for handling class imbalance effectively?

A healthcare company is using Amazon SageMaker Clarify to ensure its AI model does not introduce bias when predicting patient readmission risk. They want to analyze whether income level disproportionately affects predictions across demographic groups. Which metric should they use to assess bias?

A financial services company is developing a fraud detection system that requires an end-to-end ML pipeline including data preprocessing, feature engineering, model training, tuning, and deployment. The company also wants to ensure the workflow is scalable, version-controlled, and easy to monitor. Which AWS service is best suited for managing and automating this ML workflow?

A retail company has deployed an ML-powered recommendation engine on Amazon SageMaker endpoints. To ensure fairness and prevent bias drift, the company wants to monitor how predictions change over time for different customer groups. Which solution is the most effective for monitoring bias drift?

A data scientist is working on an ML model to predict rental prices of commercial office spaces. The dataset includes:
• Building Type and Year Constructed
• City Name
• Building Size in Square Meters
The data requires feature engineering before training. Which transformation method should be applied to each feature?

A logistics company is using ML to optimize delivery routes based on real-time traffic data. The model must provide fast responses but also be cost-effective. The company has varying request loads throughout the day. Which deployment strategies best balance performance, cost, and scalability? (Select two)

Which AWS service is used to store, share, and manage input features for ML models during training and inference?

A data scientist is building an ML model to predict loan defaults. Since the dataset is highly imbalanced, accuracy alone is not a reliable metric. Which evaluation techniques should be prioritized? (Select two)

A deep learning model is being trained on X-rays and MRIs to detect diseases. It shows high accuracy on training data but poor validation performance, indicating overfitting. Which two techniques will most effectively improve generalization?

An ML engineer is training a deep learning model on a massive dataset. The training job can tolerate interruptions and must be cost-efficient. Which approach minimizes cost while ensuring successful completion?

A fraud detection model’s accuracy has declined over time due to changes in user behavior. The company suspects data drift or model drift. Which two actions should be taken?

Which approach aligns with defense-in-depth security for generative AI applications on AWS?

A financial services company is training an ML model to detect fraudulent transactions. The dataset is highly imbalanced. Which Amazon SageMaker built-in algorithm is best suited for this problem?

A banking institution requires a manual approval process before deploying a new ML model to production endpoints. Which AWS service provides a manual approval mechanism in the ML pipeline?

A healthcare company is building a predictive model to identify high-risk patients for hospital readmission. The dataset includes both categorical (e.g., diagnosis type) and numerical features (e.g., days in hospital) and suffers from class imbalance. Which AWS service provides an automated, low-overhead solution for balancing the dataset and preparing features for ML training?

A telecommunications company wants to predict network equipment failures using historical sensor data, such as temperature, pressure, and error rates. Which Amazon SageMaker built-in algorithm is best suited for this predictive maintenance task?

A retail company trained an ML model to detect damaged products in warehouse images. The model performed well in training but failed in production due to lighting and image quality variations across different locations. Which approach requires the least effort to improve the model’s robustness?

A tech startup is developing a recommendation system for its mobile app. The company has a small engineering team and needs cost-effective, scalable infrastructure. Which AWS services provide an automated, scalable, and maintainable ML pipeline?

A financial services company deployed a credit risk prediction model but found that its AUC score was below business requirements. Which two approaches will most likely improve model performance?

A fraud detection model deployed by a bank is performing worse over time. The team suspects model drift due to changes in fraud patterns. Which two actions should the company take?

A data science team at a tech company uses SageMaker notebooks for ML development. They need consistent, centralized access control for datasets stored in Amazon S3. Which solution ensures secure and centralized access management?

A biotech company wants to deploy a custom TensorFlow model with specialized dependencies in Amazon SageMaker. Which approach best supports custom containers for training and inference?

A financial institution is struggling to build an accurate credit risk prediction model. The team wants to combine different models to improve overall performance. Which ensemble learning technique is most suitable?

An e-commerce company is training recommendation models on large datasets. The training jobs are sporadic but require significant compute power. Which AWS strategy is best for balancing cost efficiency and performance?

Which of the following aligns with defense-in-depth security for generative AI applications on AWS?

A retail company is building a customer behavior prediction model. The team wants to ensure feature consistency across training and inference. Which AWS service best meets this requirement?

A healthcare company stores ML models in Amazon SageMaker Model Registry. The company needs cross-account access to centralize model governance. Which solution enables secure, cross-account model sharing?

A financial services company deployed an ML-based fraud detection model using Amazon SageMaker Model Monitor to track accuracy over time. After several months, the model’s performance dropped despite no changes being made. What is the most likely cause of this performance decline?

A logistics company built an ML model to predict delivery time for shipments. The model’s output is a continuous numerical value representing delivery time in hours. Which evaluation metric is most suitable for measuring the model’s accuracy?

An e-commerce company uses an ML model to generate real-time product recommendations. The system experiences high traffic spikes during flash sales. Which scaling policy best handles variable traffic demand while optimizing costs?

A healthcare startup is training an AI model to classify X-ray images. The dataset includes millions of high-resolution images stored in Amazon S3, and efficient data access is required for SageMaker training. Which SageMaker input mode is best suited for handling large-scale image datasets?

A global e-commerce company built a custom sentiment analysis model in Amazon Comprehend (Account A). Now, it needs to share the model with Account B securely. What is the best way to achieve this?

A telecom company is training a customer churn prediction model. The dataset includes purchase history, support interactions, and subscription details. The company needs an algorithm that can handle feature dependencies and class imbalance. Which Amazon SageMaker built-in algorithm is best suited for this task?

A fintech company is developing an ML-based credit scoring system. The model needs to be retrained regularly as new customer data becomes available. Which AWS service provides an automated, scalable workflow for retraining and deployment?

A financial institution requires a manual review process before deploying ML models into production. Compliance regulations demand that only approved models be deployed. Which AWS service enables manual approval in the ML deployment pipeline?

A retail company is training an ML model using Amazon SageMaker on a large dataset. The training job can tolerate interruptions. Which AWS service provides the most cost-effective training solution?

A healthcare company is developing ML models across multiple AWS accounts. The models must be centrally managed, versioned, and accessible across teams. Which AWS service best meets these requirements?

A SaaS company wants to automate the deployment of ML models in Amazon SageMaker while continuously monitoring performance. Which AWS services should they use? (Select two)

A marketing analytics company is building a customer segmentation model. The dataset contains structured and unstructured data stored in Amazon S3. Which AWS service provides the most efficient solution for feature transformation?

A bank uses an ML model to predict loan default risk. Regulators require that the model does not discriminate against protected demographic groups. Which AWS service helps detect and mitigate bias in the model?

A marketing agency wants to deploy an NLP-based sentiment analysis model to analyze customer feedback efficiently. The company prefers minimal training time and easy deployment and has access to Amazon Bedrock and SageMaker JumpStart. Which approach is the best choice for fine-tuning an existing pre-trained NLP model?

A healthcare company wants to build a predictive ML model to assess patient readmission risk. The workflow includes data preprocessing, model training, hyperparameter tuning, deployment, and continuous retraining. Which AWS solution provides scalability, automation, and minimal manual intervention?

A financial institution requires manual approvals before deploying an ML model for credit risk assessment. The solution must include secure data storage, model version control, and governance compliance. Which solution best meets these requirements?

A fintech company is deploying an AI-powered fraud detection system and needs to securely manage model versions and deployments while ensuring data isolation for different ML workflows. Which AWS service offers the most efficient solution with minimal operational overhead?

An e-commerce company is building an ML model to predict whether a customer will make a purchase. The dataset is highly imbalanced, with only a small percentage of positive outcomes. Which two evaluation metrics should be prioritized?

A retail company wants to scale a SageMaker model that provides real-time product recommendations. During peak shopping events, traffic increases dramatically. Which scaling strategy is best for handling variable traffic efficiently?

A healthcare company is training an ML model to classify medical images stored in Amazon S3. The dataset is terabytes in size, and efficient I/O access is required. Which SageMaker input mode is best suited for this scenario?

A global e-commerce company developed a custom sentiment analysis model in Amazon Comprehend (Account A) and wants to share it securely with Account B. Which AWS method ensures efficient cross-account model sharing?

A subscription-based company is building an ML model to predict customer churn. The dataset is highly imbalanced, and feature interactions need to be considered. Which Amazon SageMaker built-in algorithm is best suited for this task?

A logistics company built an ML model to predict delivery time in hours. Which metric is best for evaluating continuous numerical predictions?

An e-commerce company needs to automate the deployment of an AI-driven recommendation engine while ensuring inter-service communication. Which approach best supports Infrastructure as Code (IaC)?

A media streaming company runs AWS Glue jobs to process metadata for ML models in SageMaker Pipelines. Which solution provides seamless integration with minimal overhead?

A healthcare startup is building an ML model to predict disease risk. The dataset is small and imbalanced, with missing values. Which first step best determines ML feasibility?

A data science team is developing a customer lifetime value (CLV) prediction model and is using Amazon SageMaker’s automatic hyperparameter tuning. The team initially used Random Search, but some trials are not converging effectively. Which adjustment is most likely to improve tuning efficiency?

A marketing analytics company is developing a customer segmentation model. The dataset consists of categorical (e.g., customer tier) and numerical (e.g., purchase amount) features and is stored in Amazon S3 and an on-premises PostgreSQL database. Which AWS service provides the most efficient preprocessing solution with minimal overhead?

A company stores millions of rows of training data in Amazon S3 and needs an interactive, visual tool for data selection, cleansing, and feature exploration. Which AWS service is best suited for this requirement?

A weather research organization collects data from IoT sensors, stores it in Amazon S3, and wants to run ad-hoc queries on the dataset based on an observation date column. Which solution provides the most efficient querying with minimal operational overhead?

A retail company has deployed an ML model on an Amazon SageMaker endpoint to forecast product demand. The company needs to log and monitor API calls and receive notifications when requests exceed a specific threshold. Which solution meets these requirements?