Here’s a refined and updated set of 50 key exam-style questions (with answers) focused on AWS Generative AI, especially targeting the AWS Certified AI Practitioner (AIF-C01) exam. This set aligns with the official exam domains and emphasizes generative AI concepts, AWS services, and real-world applications. These questions are a high-impact subset, not the full 100, but they zero in on the most exam-relevant content.
What is a token in generative AI?
Answer: A unit of text processed by the model. AWS in Plain English
Which of the following is NOT a use case for generative AI?
A) Image generation
B) Summarization
C) Data encryption
D) Code generation
Answer: C) Data encryption. AWS in Plain English
What does “hallucination” mean in generative AI?
Answer: An incorrect or fabricated output presented as fact. AWS in Plain English
Which AWS service is specifically designed for building generative AI applications?
Answer: Amazon Bedrock. AWS in Plain English
What is a foundation model?
Answer: A large pre-trained model adaptable for various tasks. AWS in Plain English
Which of these is NOT a stage in the foundation model lifecycle?
A) Data selection
B) Pre-training
C) Deployment
D) Marketing
Answer: D) Marketing. AWS in Plain English
What does prompt engineering involve?
Answer: Crafting effective input prompts to guide model outputs. AWS in Plain English
What is chunking in generative AI?
Answer: Breaking down large inputs into smaller, manageable pieces. AWS in Plain English
What is a multi-modal model?
Answer: A model that can work with multiple data types—text, images, audio. AWS in Plain English
What is a diffusion model commonly used for?
Answer: Generating images. AWS in Plain English
Which AWS service offers conversational AI?
Answer: Amazon Q. AWS in Plain EnglishWikipedia
Which AWS services support fine-tuning or testing foundation models?
Answer: Amazon Bedrock, Amazon SageMaker JumpStart, PartyRock (Bedrock Playground). AWS Static
What cost model is relevant to AWS generative AI services?
Answer: Token-based pricing. AWS in Plain EnglishAWS Static
Name an advantage of AWS generative AI infrastructure.
Answer: Security, compliance, and accelerated speed-to-market. AWS Static
What is Amazon Q?
Answer: A generative AI-powered enterprise chatbot for tasks like cloud app troubleshooting and document summarization. Wikipedia
What is Retrieval-Augmented Generation (RAG)?
Answer: A technique that augments responses with retrieval from external knowledge sources—commonly used with Bedrock. AWS Static
Which AWS services can store embeddings for generative AI?
Answer: Amazon OpenSearch, Neptune, Aurora, DocumentDB, RDS for PostgreSQL. AWS Static
Which customization approaches for foundation models involve higher cost trade-offs?
Answer: Pre-training and fine-tuning. AWS Static
How do inference parameters like “temperature” affect model output?
Answer: They control creativity vs determinism—higher temperature increases randomness. (Derived from fundamentals.) AWS Static
What are important model selection criteria for foundation models?
Answer: Latency, cost, multi-lingual capabilities, model size, complexity, customization needs. AWS Static
How do AI, ML, and Deep Learning relate?
Answer: ML is a subset of AI. Deep learning is a subset of ML. MediumAWS Static
Name the three types of machine learning.
Answer: Supervised, unsupervised, and reinforcement learning. AWS Static
What’s included in an ML development pipeline?
Answer: Data collection, preprocessing, feature engineering, training, evaluation, deployment, monitoring. AWS Static
Mention two AWS services integral to ML operations (MLOps).
Answer: SageMaker Model Monitor, SageMaker Data Wrangler, Feature Store. AWS Static
What are some model performance metrics?
Answer: Accuracy, AUC, F1 score. AWS Static
What are core limitations of generative AI?
Answer: Hallucinations, inaccuracy, lack of interpretability, nondeterminism. AWS Static
What factors should influence generative AI model selection?
Answer: Model capabilities, constraints, performance, compliance, business fit. AWS Static
Listing examples of responsible AI considerations.
Answer: Bias mitigation, fairness, transparency, explainability, governance frameworks. AWS Static
Give one AWS benefit for generative AI compliance.
Answer: Ensures security and regulatory compliance via built-in AWS controls. AWS Static
Security responsibilities when using AWS AI services?
Answer: Shared responsibility model—AWS secures infrastructure, users secure access and data. Derived from AWS certification scope. AWS Static