A company has developed a chatbot capable of answering natural language questions using images. The company wants to ensure the chatbot does not return inappropriate or unwanted images to users. Which solution meets this requirement?

B. Retrain the model using a generic public dataset.

D. Implement automated integration of user feedback.

Which AWS service or feature enables AI development teams to rapidly deploy and use foundation models (FMs) within their VPC?

C. PartyRock, the Amazon Bedrock Playground

A lending company is building a generative AI–based solution to offer personalized discounts to loan applicants based on specific eligibility criteria. The company wants to build and use the AI model responsibly and minimize potential bias that could negatively impact certain customers. Which actions should the company take to meet these requirements? (Choose two.)

C. Evaluate model behavior to provide transparency to stakeholders.

A. Detect imbalances or disparities in the data.

B. Ensure the model runs frequently.

C. Evaluate model behavior to provide transparency to stakeholders.

D. Use Recall-Oriented Understudy for Gisting Evaluation (ROUGE) to ensure the model achieves 100% accuracy.

E. Ensure inference latency remains within acceptable limits.

A company wants to use Amazon SageMaker to build machine learning models. It needs to share and manage variables required for model development across multiple teams. Which SageMaker feature satisfies this requirement?

A. Amazon SageMaker Feature Store

B. Amazon SageMaker Data Processing tools

D. Amazon SageMaker Model Cards

A company is developing an application that needs to generate synthetic data based on existing data. Which type of model can the company use to meet this requirement?

A. Generative Adversarial Network (GAN)

C. Residual Neural Network (ResNet)

A company is using few-shot prompting with a foundation model hosted on Amazon Bedrock. The model currently uses 10 examples in its prompt. The model is invoked once per day and performs well. The company wants to reduce monthly costs. Which solution satisfies these requirements?

B. Reduce the number of tokens in the prompt.

A. Fine-tune the model to customize it.

B. Reduce the number of tokens in the prompt.

C. Increase the number of tokens in the prompt.

A company wants to visualize the total sales revenue of its best-selling product across all retail locations over the past 12 months. Which AWS solution should the company use to automate chart generation?

C. Amazon Q in Amazon QuickSight

In the context of generative AI models, what are tokens?

A. Tokens are the fundamental units of input and output processed by generative AI models; they can represent words, subwords, or other linguistic units.

B. In generative AI models, tokens are mathematical representations of words or concepts.

C. Tokens refer to pre-trained weights in generative AI models that are fine-tuned for specific tasks.

D. Tokens are specific prompts or instructions given to a generative AI model to produce output.

AWS AIF-C01 Chapter 2 Free Practice Test 2025 — AI Practitioner Exam Questions

AWS AIF-C01 Chapter 2 Practice Questions (100 Questions)

Practice questions for the AWS AIF-C01 (AI Practitioner) exam, Chapter 2.

Q1. A company has developed a chatbot capable of answering natural language questions using images. The company wants to ensure the chatbot does not return inappropriate or unwanted images to users. Which solution meets this requirement?
- A. Implement a moderation API.
- B. Retrain the model using a generic public dataset.
- C. Perform model validation.
- D. Implement automated integration of user feedback.
View question →
Q2. Which AWS service or feature enables AI development teams to rapidly deploy and use foundation models (FMs) within their VPC?
- A. Amazon Personalize
- B. Amazon SageMaker JumpStart
- C. PartyRock, the Amazon Bedrock Playground
- D. Amazon SageMaker endpoints
View question →
Q3. A lending company is building a generative AI–based solution to offer personalized discounts to loan applicants based on specific eligibility criteria. The company wants to build and use the AI model responsibly and minimize potential bias that could negatively impact certain customers. Which actions should the company take to meet these requirements? (Choose two.)
- A. Detect imbalances or disparities in the data.
- B. Ensure the model runs frequently.
- C. Evaluate model behavior to provide transparency to stakeholders.
- D. Use Recall-Oriented Understudy for Gisting Evaluation (ROUGE) to ensure the model achieves 100% accuracy.
- E. Ensure inference latency remains within acceptable limits.
View question →
Q4. A company wants to use Amazon SageMaker to build machine learning models. It needs to share and manage variables required for model development across multiple teams. Which SageMaker feature satisfies this requirement?
- A. Amazon SageMaker Feature Store
- B. Amazon SageMaker Data Processing tools
- C. Amazon SageMaker Clarify
- D. Amazon SageMaker Model Cards
View question →
Q5. A company is developing an application that needs to generate synthetic data based on existing data. Which type of model can the company use to meet this requirement?
- A. Generative Adversarial Network (GAN)
- B. XGBoost
- C. Residual Neural Network (ResNet)
- D. WaveNet
View question →
Q6. A company wants to classify human genes into 20 categories based on genetic features. The company requires a machine learning method that can explain how the model's internal mechanisms influence its output. Which machine learning method satisfies these requirements?
- A. Decision tree
- B. Linear regression
- C. Logistic regression
- D. Neural network
View question →
Q7. A company is using few-shot prompting with a foundation model hosted on Amazon Bedrock. The model currently uses 10 examples in its prompt. The model is invoked once per day and performs well. The company wants to reduce monthly costs. Which solution satisfies these requirements?
- A. Fine-tune the model to customize it.
- B. Reduce the number of tokens in the prompt.
- C. Increase the number of tokens in the prompt.
- D. Use provisioned throughput.
View question →
Q8. A company wants to visualize the total sales revenue of its best-selling product across all retail locations over the past 12 months. Which AWS solution should the company use to automate chart generation?
- A. Amazon Q in Amazon EC2
- B. Amazon Q Developer
- C. Amazon Q in Amazon QuickSight
- D. Amazon Q in AWS Chatbot
View question →
Q9. In the context of generative AI models, what are tokens?
- A. Tokens are the fundamental units of input and output processed by generative AI models; they can represent words, subwords, or other linguistic units.
- B. In generative AI models, tokens are mathematical representations of words or concepts.
- C. Tokens refer to pre-trained weights in generative AI models that are fine-tuned for specific tasks.
- D. Tokens are specific prompts or instructions given to a generative AI model to produce output.
View question →
Q10. A company wants to use a large language model (LLM) on Amazon Bedrock for sentiment analysis. The company wants to understand how much information a prompt can contain. On which factor will the company’s decision be based?
- A. Temperature
- B. Context window
- C. Batch size
- D. Model size
View question →
Q11. When working with foundation models (FMs), which option is beneficial for performing continual pretraining?
- A. Helps reduce model complexity
- B. Model performance improves over time
- C. Reduces training time required
- D. Optimizes model inference time
View question →
Q12. A company is building a machine learning model. It has collected raw data and is analyzing it by creating correlation matrices, computing statistical summaries, and visualizing distributions. At which stage of the machine learning pipeline is the company currently operating?
- A. Data preprocessing
- B. Feature engineering
- C. Exploratory data analysis (EDA)
- D. Hyperparameter tuning
View question →
Q13. A marketing team is using a large language model (LLM) to generate content for a campaign. The generated content appears plausible and factually consistent—but is inaccurate. What issue is the LLM encountering?
- A. Data leakage
- B. Hallucination
- C. Overfitting
- D. Misalignment
View question →
Q14. A company wants to develop an educational game where users need to answer the following question: 'A jar contains six red, four green, and three yellow balls. What is the probability of randomly selecting a green ball from the jar?' Which solution meets these requirements while minimizing cost?
- A. Use supervised learning to create a regression model that predicts the probability.
- B. Use reinforcement learning to train a model that returns the probability.
- C. Use code with simple rules and calculations to compute the probability.
- D. Use unsupervised learning to create a model that estimates probability density.
View question →
Q15. A healthcare company has deployed a disease detection model on Amazon Bedrock. To comply with privacy policies, the company wants to prevent patient personally identifiable information (PII) from appearing in model responses. Additionally, the company wants to receive notifications if a policy violation occurs. Which solution satisfies these requirements?
- A. Use Amazon Macie to scan model outputs for sensitive data and set alarms for potential violations.
- B. Configure AWS CloudTrail to monitor model responses and create alarms for any detected personal information.
- C. Use Amazon Bedrock Guardrails to filter content and configure Amazon CloudWatch alarms to notify on policy violations.
- D. Implement Amazon SageMaker Model Monitor to detect data drift and receive alerts when model quality degrades.
View question →
Q16. A company experiences database errors causing missing words in documents. It wants to build a machine learning model that suggests plausible words to fill in missing text. Which type of model best fits this requirement?
- A. Topic modeling
- B. Clustering model
- C. Prescriptive machine learning model
- D. BERT-based model
View question →
Q17. An AI practitioner maintains an animal photo database and wants to automatically identify and classify animals in photos without human intervention. Which strategy satisfies these requirements?
- A. Object detection
- B. Anomaly detection
- C. Named entity recognition
- D. Painting
View question →
Q18. A company wants to build a large language model (LLM) application using Amazon Bedrock and customer data stored in Amazon S3. Its security policy mandates that each team can access only its own team’s customer data. Which solution satisfies these requirements?
- A. Create an Amazon Bedrock custom service role for each team, granting access only to that team’s customer data.
- B. Create a custom service role with Amazon S3 permissions and require teams to specify customer names in every Amazon Bedrock request.
- C. Edit personal data in Amazon S3 and update the S3 bucket policy to allow team access to customer data.
- D. Create an Amazon Bedrock role with full Amazon S3 access and assign each team an IAM role restricted to its customer folder.
View question →
Q19. A company wants to use artificial intelligence to protect its applications from threats. The AI solution must determine whether an IP address originates from a suspicious source. Which solution satisfies this requirement?
- A. Build a speech recognition system.
- B. Create a natural language processing (NLP) named entity recognition system.
- C. Develop an anomaly detection system.
- D. Create a fraud prediction system.
View question →
Q20. A financial institution is developing an AI application using Amazon Bedrock and hosting it within a Virtual Private Cloud (VPC). To meet regulatory compliance standards, the VPC must not permit any internet-bound traffic. Which AWS service or feature satisfies this requirement?
- A. AWS PrivateLink
- B. Amazon Macie
- C. Amazon CloudFront
- D. Internet Gateway
View question →
Q21. A company has built a deep learning model for object detection and deployed it into production. When the model analyzes an image to identify objects, which AI process occurs?
- A. Training
- B. Inference
- C. Model deployment
- D. Bias correction
View question →
Q22. A social media company wants to use a large language model (LLM) for content moderation. The company wants to evaluate whether the LLM's output contains bias, including potential discrimination against specific demographic groups or individuals. Which data source should the company use to evaluate the LLM's output with the least required management overhead?
- A. User-generated content
- B. Moderation logs
- C. Content moderation guidelines
- D. Benchmark dataset
View question →
Q23. Which feature of Amazon OpenSearch Service enables companies to build vector database applications?
- A. Integration with Amazon S3 for object storage
- B. Support for geospatial indexing and queries
- C. Scalable index management and nearest neighbor search capabilities
- D. Real-time analysis of streaming data
View question →
Q24. Which strategy is used to evaluate the accuracy of a foundation model (FM) in an image classification task?
- A. Calculate the total resource cost incurred by the model.
- B. Evaluate model accuracy against a predefined benchmark dataset.
- C. Count the number of layers in the neural network.
- D. Assess color accuracy of images after model processing.
View question →
Q25. A company wants to use large language models (LLMs) from Amazon Bedrock to develop a chat interface for its product documentation, which is stored as PDF files. Which solution meets these requirements in the most cost-effective way?
- A. Use prompt engineering to include one PDF file as context in the user prompt when submitting to Amazon Bedrock.
- B. Use prompt engineering to include all PDF files as context in the user prompt when submitting to Amazon Bedrock.
- C. Fine-tune the model using all PDF documents via Amazon Bedrock, then use the fine-tuned model to process user prompts.
- D. Upload the PDF documents to an Amazon Bedrock knowledge base. When users submit prompts to Amazon Bedrock, leverage the knowledge base to provide contextual information.
View question →
Q26. Which option represents a use case for generative AI models?
- A. Improving cybersecurity using an intrusion detection system.
- B. Generating realistic images from text descriptions for digital marketing.
- C. Enhancing database performance using optimized indexes.
- D. Analyzing financial data to predict stock market trends.
View question →
Q27. A company uses a foundation model (FM) from Amazon Bedrock to develop an enterprise search tool. The company wants to adapt the model using its own data to improve accuracy. Which strategy successfully adapts the model?
- A. Provide labeled data with prompt and completion fields.
- B. Prepare the training dataset as a .txt file containing multiple lines of CSV-formatted content.
- C. Purchase preconfigured throughput for Amazon Bedrock.
- D. Train the model using journals and textbooks.
View question →
Q28. A company is using a pre-trained large language model (LLM) to build a chatbot for product recommendations. The company wants the LLM’s output to be concise and follow a specific language style. Which solution ensures the LLM’s response quality meets the company’s expectations?
- A. Adjust the prompt.
- B. Select a different-sized LLM.
- C. Increase temperature.
- D. Increase Top-K value.
View question →
Q29. A company is using the Generative AI Security Scope Matrix to assess its solution’s security accountability. It has already defined several solution scopes based on this matrix. Which solution scope enables the company to assume the greatest degree of security accountability?
- A. Using a third-party enterprise application embedded with generative AI functionality.
- B. Building an application using an existing third-party generative AI foundation model (FM).
- C. Optimizing an existing third-party generative AI foundation model (FM) by fine-tuning it with business-specific data.
- D. Building and training a generative AI model from scratch using customer-owned, domain-specific data.
View question →
Q30. A company is developing a call center application and wants to gain insights from customer calls. The company wants to analyze and extract key information from customer call audio. Which solution meets these requirements?
- A. Use Amazon Lex to build a conversational chatbot.
- B. Use Amazon Transcribe to transcribe call recordings.
- C. Use Amazon SageMaker Model Monitor to extract information from call recordings.
- D. Use Amazon Comprehend to create classification labels.
View question →
Q31. A company needs to build its own large language model (LLM) exclusively using its private data. The company is concerned about the environmental impact of the training process. Which Amazon EC2 instance type minimizes environmental impact when training large language models (LLMs)?
- A. Amazon EC2 C series
- B. Amazon EC2 G series
- C. Amazon EC2 P series
- D. Amazon EC2 Trn series
View question →
Q32. A developer trained a custom model on Amazon Bedrock using a training dataset containing confidential data. The developer wants to ensure the custom model does not generate inference responses based on the confidential data. What should the developer do to prevent inference responses based on confidential data?
- A. Delete the custom model. Remove confidential data from the training dataset. Retrain the custom model.
- B. Apply dynamic data masking to mask confidential data in inference responses.
- C. Use Amazon SageMaker to encrypt confidential data in inference responses.
- D. Use AWS Key Management Service (AWS KMS) to encrypt confidential data within the custom model.
View question →
Q33. A company is developing a solution aimed at generating images for safety glasses. The solution must achieve high accuracy and minimize the risk of incorrect outputs.
- A. Use human-in-the-loop validation with Amazon SageMaker Ground Truth Plus.
- B. Perform data augmentation using an Amazon Bedrock knowledge base.
- C. Use Amazon Rekognition for image recognition.
- D. Use Amazon QuickSight for data aggregation.
View question →
Q34. A company wants to use a domain-specific model. It wishes to avoid building a model from scratch. Instead, it wants to fine-tune a pre-trained model to create a model tailored to its specific tasks.
- A. Increase the number of iterations.
- B. Use transfer learning.
- C. Decrease the number of iterations.
- D. Use unsupervised learning.
View question →
Q35. Amazon SageMaker Clarify提供曾些功能?
- A. 整合了检索增强生成(RAG)工作流程
- B. 监控生产环境中机器学习模型的质量
- C. 记录机器学习模型的关键细节
- D. 在数据准备过程中识别潜在偏差
View question →
Q36. A company wants to develop an interactive application for children that generates stories based on classic tales. The company plans to use Amazon Bedrock and must ensure the generated stories and themes are appropriate for children.
- A. Amazon Rekognition
- B. Amazon Bedrock Playground
- C. Amazon Bedrock Guardrails
- D. Amazon Bedrock Agents
View question →
Q37. A large retail company receives thousands of customer support inquiries about products daily. These inquiries require prompt resolution and responses. The company wants to deploy agents on Amazon Bedrock.
- A. Generate custom foundation models (FMs) to predict customer needs.
- B. Automate repetitive tasks and orchestrate complex workflows.
- C. Automatically invoke multiple foundation models (FMs) and aggregate results.
- D. Select foundation models (FMs) based on predefined criteria and metrics.
View question →
Q38. A law firm wants to leverage large language models (LLMs) to develop a developer application that reads legal documents and extracts key points.
- A. Build an automated named entity recognition system.
- B. Create a recommendation engine.
- C. Develop a summarization chatbot.
- D. Develop a multilingual translation system.
View question →
Q39. A company collects petabytes of unstructured data from on-premises sources and stores it in a data lake. The company wants to convert this data into structured formats so that its data scientists can perform machine learning (ML) tasks. Which service meets these requirements?
- A. Amazon Lex
- B. Amazon Rekognition
- C. Amazon Kinesis Data Streams
- D. AWS Glue
View question →
Q40. A company wants to develop a chatbot to assist customers. The chatbot must help resolve technical issues without human intervention under normal operating conditions. The company has selected a foundation model (FM) for the chatbot. The chatbot needs to generate responses consistent with the company's tone. Which solution meets these requirements?
- A. Set a low limit on the number of tokens the FM can generate.
- B. Use batch inference to process detailed responses.
- C. Experiment with and optimize prompts until the FM produces the desired responses.
- D. Set a higher value for the temperature parameter.
View question →
Q41. An educational institution is developing a question-and-answer application that uses generative AI models to explain complex concepts. The institution wants the model’s responses to automatically adapt their tone based on the questioner’s identity. The institution will provide the model with the user’s age range. Which solution satisfies these requirements with minimal implementation effort?
- A. Fine-tune the model using additional training data representing all age groups supported by the application.
- B. Add a role description in the prompt context instructing the model which age group to target in its response.
- C. Apply chain-of-thought reasoning to infer the correct tone and complexity level suitable for the user’s response.
- D. Summarize the response text based on the user’s age to deliver more concise replies to younger users.
View question →
Q42. An AI developer has built a deep learning model to classify material types in images. Now, the developer wants to evaluate the model’s performance. Which metric helps the AI developer evaluate the model’s performance?
- A. Confusion matrix
- B. Correlation matrix
- C. R² score
- D. Mean squared error (MSE)
View question →
Q43. An AI company periodically engages independent software vendors (ISVs) to assess its systems and processes. When an ISV’s compliance report becomes available, the company requires an email notification. Which AWS service can the company use to meet this requirement?
- A. AWS Audit Manager
- B. AWS Artifact
- C. AWS Trusted Advisor
- D. AWS Data Exchange
View question →
Q44. A company wants to use a large language model (LLM) on Amazon Bedrock for sentiment analysis. The company requires the LLM to produce consistent responses for identical input prompts. What inference parameter adjustment should the company make to meet these requirements?
- A. Decrease the temperature value.
- B. Increase the temperature value.
- C. Reduce the output token length.
- D. Increase the maximum generation length.
View question →
Q45. A security company is using Amazon Bedrock to host foundation models (FMs). The company wants to ensure only authorized users can access these models. It needs to identify any unauthorized access attempts to configure appropriate AWS Identity and Access Management (IAM) policies and roles for future iterations of the foundation models. Which AWS service should the company use to identify unauthorized users accessing Amazon Bedrock?
- A. AWS Audit Manager
- B. AWS CloudTrail
- C. Amazon Fraud Detector
- D. AWS Trusted Advisor
View question →
Q46. A company has petabytes of customer transaction data used for advertising campaigns. The company wants to segment customers into different tiers to enable targeted advertising and product promotion. Which method should the company adopt to meet these requirements?
- A. Supervised learning
- B. Unsupervised learning
- C. Reinforcement learning
- D. Reinforcement learning from human feedback (RLF)
View question →
Q47. A company is using Amazon SageMaker Studio notebooks to build and train machine learning models. The company stores its data in an Amazon S3 bucket. The company needs to manage the data flow from Amazon S3 to SageMaker Studio notebooks. Which solution meets this requirement?
- A. Use Amazon Inspector to monitor SageMaker Studio.
- B. Use Amazon Macie to monitor SageMaker Studio.
- C. Configure SageMaker to use a VPC with an S3 endpoint.
- D. Configure SageMaker to use S3 Glacier Deep Archive.
View question →
Q48. A healthcare company is developing a foundation model (FM) for diagnostic purposes. The company requires the model to be transparent and easily interpretable to comply with regulatory requirements. Which solution meets these requirements?
- A. Use Amazon Inspector to configure security and compliance.
- B. Use Amazon SageMaker Clarify to generate simple metrics, reports, and examples.
- C. Use Amazon Macie to encrypt and protect training data.
- D. Collect more data and use Amazon Rekognition to add custom labels to the data.
View question →
Q49. Which term describes the numerical representation of real-world objects and concepts used by artificial intelligence and natural language processing (NLP) models to improve understanding of textual information?
- A. Embeddings
- B. Tokens
- C. Model
- D. Binary files
View question →
Q50. A company is developing a chatbot to improve user experience. It is using a large language model (LLM) from Amazon Bedrock for intent detection and wants to apply few-shot learning to improve intent detection accuracy. What additional data does the company need to meet these requirements?
- A. Multiple sets of chatbot responses paired with correct user intents.
- B. Pairs of user messages and correct chatbot responses.
- C. Pairs of user messages and correct user intents.
- D. Pairs of user intents and correct chatbot responses.
View question →
Q51. A company wants to perform sentiment analysis using a large language model (LLM) on Amazon Bedrock, classifying text passages as positive or negative. Which prompt engineering strategy meets this requirement?
- A. Provide several example text passages labeled with their corresponding positive or negative sentiment in the prompt, followed by the new passage to classify.
- B. Explain sentiment analysis and how LLMs work in detail within the prompt.
- C. Provide only the new text passage to classify, without any additional context or examples.
- D. Include several unrelated task examples in the prompt, such as text summarization or question answering.
View question →
Q52. A company has a foundation model (FM) customized using Amazon Bedrock to answer customer queries about products. The company wants to validate the model’s responses to specific query types and needs to provide a validation dataset for Amazon Bedrock to use. Which AWS service meets this requirement?
- A. Amazon S3
- B. Amazon Elastic Block Store (Amazon EBS)
- C. Amazon Elastic File System (Amazon EFS)
- D. AWS Snowcone
View question →
Q53. A company has terabytes of data in its database available for business analytics. It wants to build an AI-powered application that generates SQL queries based on natural-language text input from employees who have very limited technical expertise. Which solution meets this requirement?
- A. Generative Pre-trained Transformer (GPT)
- B. Residual Neural Network
- C. Support Vector Machine
- D. WaveNet
View question →
Q54. A company wants to use a large language model (LLM) to develop a conversational AI agent. It needs to prevent the LLM from being manipulated by common prompt injection techniques, which could cause undesirable behavior or leakage of sensitive information. What action can reduce these risks?
- A. Create a prompt template to train the LLM to detect attack patterns.
- B. Increase the temperature parameter in LLM inference requests.
- C. Avoid using LLMs not listed in Amazon SageMaker.
- D. Reduce the number of input tokens when invoking the LLM.
View question →
Q55. A company wants to use Amazon Bedrock to develop an application. The company has a limited budget and prefers flexibility, without long-term commitments.
- A. On-Demand
- B. Model Customization
- C. Provisioned Throughput
- D. Spot Instances
View question →
Q56. How can a company securely use large language models (LLMs) on Amazon Bedrock?
- A. Design clear and specific prompts. Configure AWS Identity and Access Management (IAM) roles and policies with least-privilege access.
- B. Enable AWS Audit Manager to automate model evaluation jobs.
- C. Enable Amazon Bedrock automated model evaluation jobs.
- D. Use Amazon CloudWatch Logs to improve model interpretability and monitor for bias.
View question →
Q57. A company performs quarterly forecasts to optimize operations and meet anticipated demand. It uses machine learning models for these forecasts. A data scientist must author a report on the trained machine learning model to provide transparency and explainability to stakeholders.
- A. Model training code
- B. Partial Dependence Plots (PDPs)
- C. Training sample data and model convergence table
View question →
Q58. An accounting firm wants to adopt large language models (LLMs) to automate document review. The firm must act responsibly and avoid potential risks. Which measures should the firm take when developing and deploying LLMs? (Choose two.)
- A. Incorporate fairness metrics in model evaluation.
- B. Adjust the model’s temperature parameter.
- C. Modify training data to reduce bias.
- D. Avoid overfitting on training data.
- E. Apply prompt engineering techniques.
View question →
Q59. A data scientist is building a model to generate images of people in various occupations. The data scientist discovers that input data contains bias, and certain attributes influence image generation, introducing bias into the model.
- A. Data augmentation for imbalanced classes
- B. Model monitoring for class distribution
- C. Retrieval-Augmented Generation (RAG)
- D. Image watermark detection
View question →
Q60. A company wants to use a pre-trained generative AI model to create marketing content. It needs to ensure generated content aligns with its brand voice and messaging requirements.
- A. Optimize model architecture and hyperparameters to improve overall model performance.
- B. Increase model complexity by adding more layers to the architecture.
- C. Create effective prompts with clear instructions and contextual guidance to steer model output.
- D. Select a large and diverse dataset to pre-train a new generative model.
View question →
Q61. A company is developing a model to predict product prices. The model performs well on its training dataset but shows significantly degraded performance when deployed into production.
- A. Reduce the amount of data used during training.
- B. Add hyperparameters to the model.
- C. Increase the amount of data used during training.
- D. Increase model training time.
View question →
Q62. A company is developing a customer service chatbot and wants to enhance its response capability by learning from past interactions and online resources.
- A. Supervised learning using a manually curated dataset of correct and incorrect answers.
- B. Reinforcement learning with rewards for positive customer feedback.
- C. Unsupervised learning to cluster similar customer inquiries.
- D. Supervised learning using a continuously updated FAQ database.
View question →
Q63. A social media company wants to use AI to detect harmful images and video content on its platform. Which AWS service can meet this requirement with minimal operational overhead?
- A. Amazon Rekognition
- B. Amazon SageMaker
- C. Amazon Transcribe
- D. Amazon Textract
View question →
Q64. A marketing company uses a large language model (LLM). The company wants to evaluate how the LLM’s response quality changes when minor perturbations are applied to the input in a question-answering task. Which metric should the company use?
- A. Root Mean Square Error (RMSE)
- B. Area Under the ROC Curve (AUC)
- C. F1 Score
- D. Semantic Robustness
View question →
Q65. A company wants to enhance its loan approval process. The company seeks to ensure ML prediction accuracy through human review. Which AWS service can build the workflow required for human review of ML predictions?
- A. Amazon Augmented AI (Amazon A2I)
- B. Amazon Textract
- C. Amazon SageMaker Ground Truth
- D. Amazon Lex
View question →
Q66. Select the correct machine learning technique from the list below to build a voice command recognition application. Which ML technique should be selected to train a model to recognize voice commands based on labeled audio data?
- A. Supervised Learning
- B. Semi-supervised Learning
- C. Unsupervised Learning
View question →
Q67. In the ML context, what does bias refer to?
- A. Model architecture complexity
- B. Randomness or noise in training data
- C. Difference between predicted and actual values
- D. Systematic error or tendency of a model to make incorrect predictions for certain data
View question →
Q68. A company operating an online learning platform wants to build a recommendation engine that recommends relevant courses based on user interests and learning history. Which AWS service meets the requirements for building and deploying such a recommendation engine?
- A. Amazon SageMaker
- B. AWS Lambda
- C. Amazon Rekognition
- D. Amazon Comprehend
View question →
Q69. A company wants to analyze customer purchase data. The company aims to group customers to improve targeted marketing campaigns and intends to use an unlabeled dataset. Which ML technique satisfies these requirements?
- A. Unsupervised Learning
- B. Fine-tuning
- C. Supervised Learning
- D. Transfer Learning
View question →
Q70. A company collects large tabular datasets weekly for its ML workflow. The workflow can tolerate up to 7 days of latency between data ingestion and inference. Inference may take several hours. The company wants a solution that performs inference on these datasets without incurring endpoint costs when idle. Which type of inference satisfies these requirements?
- A. Real-time Inference
- B. Batch Transform
- C. Serverless Inference
- D. Asynchronous Inference
View question →
Q71. A data scientist is building an ML pipeline to train a text classification model. The data scientist has already collected the data for training. Which component of the ML lifecycle must be completed first?
- A. Model training
- B. Feature engineering
- C. Model validation
- D. Model monitoring
View question →
Q72. What is the objective of training an ML model on a large dataset?
- A. Optimize model parameter values to fit the training data
- B. Create new algorithms for ML
- C. Clean and preprocess raw data before modeling
- D. Deploy the model into production
View question →
Q73. A company is developing an ML model that requires explainability—i.e., understanding why a particular prediction was made. Which development phase supports implementing explainability techniques?
- A. Data preprocessing
- B. Model training
- C. Feature engineering
- D. Model evaluation
View question →
Q74. A law firm wants to build a solution to extract text from audio recordings. Which AWS service meets this requirement?
- A. Amazon Translate
- B. Amazon Transcribe
- C. Amazon Comprehend
- D. Amazon Rekognition
View question →
Q75. A real estate company wants to build an ML model to predict residential property sale prices based on various factors—including location, size, number of bedrooms, and amenities. Which ML technique satisfies this requirement?
- A. Classification
- B. Clustering
- C. Reinforcement learning
- D. Regression
View question →
Q76. Which type of ML involves training a model to maximize cumulative reward based on feedback received from an environment?
- A. Supervised learning
- B. Unsupervised learning
- C. Semi-supervised learning
- D. Reinforcement learning
View question →
Q77. A company has deployed a computer vision model. It wants to continuously monitor model performance and detect any drift or degradation over time. Which AWS service or feature satisfies these requirements?
- A. Amazon SageMaker Clarify
- B. Amazon SageMaker Model Monitor
- C. Amazon Augmented AI (Amazon A2I)
- D. AWS CloudTrail
View question →
Q78. A company wants to create an ML solution to help sales representatives decide whether to prepare a sales proposal for a specific customer. The model will consider input data about the customer and proposal specifications. The output will be either "Approved" or "Rejected" for each sales proposal. The company already has historical labeled records. Which ML technique satisfies these requirements?
- A. Classification
- B. Regression
- C. Clustering
- D. Dimensionality reduction
View question →
Q79. A company wants a solution to convert text to speech. Which AWS service will meet this requirement with the least operational overhead?
- A. Amazon SageMaker
- B. Amazon Comprehend
- C. Amazon Polly
- D. Amazon Transcribe
View question →
Q80. An ML engineer has trained a deep neural network model on a large dataset. What is the process called when using the trained model to make predictions on new, previously unseen data?
- A. Training
- B. Validation
- C. Inference
- D. Feature Engineering
View question →
Q81. Select the correct AWS service from the list below for the following use case. Only one AWS service should be selected. Access pre-trained generative AI models via API without managing any infrastructure.
- A. Amazon SageMaker
- B. Amazon Bedrock
- C. PartyRock, an Amazon Bedrock Playground
View question →
Q82. A company needs to identify a generative AI model capable of interpreting image content. Which type of model satisfies these requirements?
- A. Large Language Model (LLM)
- B. Diffusion Model
- C. Multimodal Model
- D. Natural Language Processing (NLP) Model
View question →
Q83. Select the correct learning method(s) from the list below for the following use case. One or more learning methods may be selected. Building a product recommendation system that recommends similar products based on examples.
- A. Zero-shot Learning
- B. One-shot Learning
- C. Few-shot Learning
View question →
Q84. A data scientist wants an AI coding assistant that integrates with Amazon SageMaker Studio to provide code suggestions and improve coding productivity. Which AWS service or application satisfies this requirement?
- A. PartyRock, an Amazon Bedrock Playground
- B. AWS Inferentia
- C. Amazon Q Developer
- D. Amazon Bedrock
View question →
Q85. A company built an AI chatbot to answer employee questions. Some employees report that generated responses are nonsensical relative to the questions asked. Which concept does this describe?
- A. Underfitting
- B. Hallucination
- C. Overfitting
- D. Bias
View question →
Q86. A company is exploring foundation models (FMs) to build a custom solution for domain-specific data. The company does not want to manage underlying infrastructure and requires intellectual property (IP) protection features. Which solution satisfies these requirements?
- A. Using Hugging Face FMs available in Amazon SageMaker
- B. Using FMs available in Amazon SageMaker JumpStart
- C. Using FMs available in AWS Deep Learning Containers
- D. Using Amazon Titan FMs available in Amazon Bedrock
View question →
Q87. A company wants to implement a solution for semantic search in PDF documents that contain images. The company plans to extract each image from the PDF, analyze the images to obtain contextual information, and then use this information for semantic search. Which solution meets these requirements?
- A. Use tokenization
- B. Use embeddings
- C. Use an LLM
- D. Use prompt engineering
View question →
Q88. Which AWS service enables building conversational interfaces using speech and text in applications?
- A. Amazon Polly
- B. Amazon Lex
- C. Amazon Comprehend
- D. Amazon Personalize
View question →
Q89. A company wants to deploy, fine-tune, and evaluate pre-trained foundation models (FMs). The company wishes to use popular model hubs such as Hugging Face as its generative AI platform. Which solution meets this requirement with minimal operational overhead?
- A. Amazon SageMaker Model Registry
- B. Amazon SageMaker JumpStart
- C. Amazon Augmented AI (Amazon A2I)
- D. Amazon Q
View question →
Q90. A company wants to use AI to improve business efficiency. The company wants to begin exploring foundation models (FMs) suitable for its specific use case. Which solution meets these requirements with minimal operational effort?
- A. Fine-tune an FM using Amazon SageMaker JumpStart
- B. Develop a new FM
- C. Use an established natural language processing (NLP) model deployed on Amazon EC2 instances
- D. Select an FM using Amazon Bedrock
View question →
Q91. A company wants to simplify the creation of business intelligence (BI) reports using natural language queries. The company wants a no-code solution. Which AWS service or feature meets these requirements?
- A. Amazon Bedrock agents
- B. Amazon Q Business
- C. Amazon SageMaker JumpStart
- D. Amazon Q in QuickSight
View question →
Q92. What is a limitation of generative AI applications when solving business problems?
- A. Lack of scalability
- B. Inability to handle large datasets
- C. Generation of biased outputs
- D. Limited business use cases
View question →
Q93. A company wants to provide employees with a chatbot capable of answering questions about internal confidential information. This information currently resides across hundreds of company documents. Which AWS managed service should the company use to implement the chatbot solution with minimal operational overhead?
- A. Amazon Q Business
- B. Amazon SageMaker JumpStart
- C. PartyRock, an Amazon Bedrock Playground
- D. Amazon Bedrock
View question →
Q94. A financial company has pre-trained its own large language model (LLM) on proprietary data. Which AWS service can the company use to deploy the LLM?
- A. Amazon SageMaker
- B. Amazon Comprehend
- C. Amazon Bedrock
- D. Amazon Lex
View question →
Q95. A company wants to add audio generation capabilities to its book summarization application. Which AWS service requires the least operational effort to meet this requirement?
- A. Amazon Comprehend
- B. Amazon SageMaker
- C. Amazon Bedrock
- D. Amazon Polly
View question →
Q96. Which AWS service is suitable for storing vector embeddings and performing semantic search?
- A. Amazon OpenSearch Service
- B. Amazon DynamoDB
- C. Amazon ElastiCache
- D. MySQL database on Amazon EC2 instances
View question →
Q97. A company wants to evaluate the quality of machine-generated summary text relative to reference text. Which metric satisfies these requirements?
- A. Semantic Robustness
- B. Accuracy
- C. Mean Absolute Error (MAE)
- D. Recall-Oriented Understudy for Gisting Evaluation (ROUGE)
View question →
Q98. A generative AI expert is conducting prompt engineering. The expert provides a prompt describing the desired output but does not include examples. Which technique does this describe?
- A. Chain-of-Thought
- B. Few-Shot Learning
- C. Zero-Shot Learning
- D. One-Shot Learning
View question →
Q99. A company wants to enhance the capabilities of a foundation model (FM) on Amazon Bedrock for the healthcare domain. The company possesses unlabeled clinical diagnosis data that does not contain personally identifiable information (PII). Which solution satisfies these requirements?
- A. Continued Pre-training
- B. Fine-tuning
- C. Prompt Engineering
- D. Implementing an Amazon Bedrock Agent
View question →
Q100. A data scientist is using a generative AI model to generate text. What is the purpose of setting inference parameters (e.g., response length and stop sequences)?
- A. To control the model’s generated output and ensure it meets required criteria
- B. To fine-tune model parameters for improved performance
- C. To specify the input data format for the model
- D. To select the pre-trained model for inference
View question →