Amazon Web Services · Free Practice Questions · Last reviewed May 2026
30real exam-style questions organised by domain, each with the correct answer highlighted and a plain-English explanation of why it's right — and why the others are wrong.
28% of exam · 6 sample questions below
A marketing firm uses Amazon Bedrock to generate ad copy. They notice that the generated text often includes factual inaccuracies about their products. Which technique would most effectively reduce these inaccuracies?
Implement Retrieval-Augmented Generation (RAG) with a product knowledge base.
Retrieval-Augmented Generation grounds each response in retrieved product facts, so the model conditions on authoritative content rather than relying solely on parametric memory. This directly targets the factual inaccuracies described, since the product knowledge base supplies verified details at inference time, satisfying the accuracy constraint in the stem.
Use longer, more detailed prompts.
Increase the temperature parameter to 0.9.
Fine-tune the model on a dataset of previous ad copies.
A developer is using Amazon Bedrock to build a chatbot that answers customer queries. The chatbot must only respond based on the provided company documentation. Which approach best meets this requirement?
Use prompt engineering to instruct the model to only use documentation.
Use a RAG architecture with the company documentation as the knowledge base.
RAG constrains responses by retrieving passages from the company documentation and supplying them as context, so answers derive from that corpus rather than the model's parametric knowledge. This satisfies the requirement to answer only from provided documentation.
Fine-tune a foundation model on the company documentation.
Use a text classification model to filter responses.
A financial services company is deploying a foundation model to analyze customer sentiment from call transcripts. The model outputs must be consistent and deterministic for auditing purposes. Which parameter configuration should the company use?
Set temperature to 0.1 and top_p to 0.9.
Set temperature to 0.7 and top_p to 1.0.
Set temperature to 0.5 and top_p to 0.5.
Set temperature to 0 and top_p to 1.
Setting temperature to 0 makes the model select the highest-probability token at each step, eliminating the random sampling that produces run-to-run variation. Keeping top_p at 1 disables nucleus filtering, so it cannot reintroduce randomness. This satisfies the audit requirement for consistent, deterministic sentiment outputs from identical transcripts.
An e-commerce company is using a foundation model to generate product descriptions. They want to reduce costs by caching frequently requested descriptions. Which AWS service should they use to implement a cache?
Amazon CloudFront
Amazon DynamoDB
Amazon S3
Amazon ElastiCache
Amazon ElastiCache provides an in-memory cache, delivering sub-millisecond latency for repeatedly requested product descriptions, which directly satisfies the stem's cost-reduction constraint by offloading duplicate foundation model inferences. Unlike persistent stores, its volatile, RAM-based architecture suits transient cached text, avoiding repeated compute charges.
A company wants to use a foundation model to automatically moderate user-generated content. The model must filter out inappropriate content with high accuracy. Which Amazon service is best suited for this task?
Amazon Translate
Amazon Rekognition
Amazon Polly
Amazon Comprehend
Amazon Comprehend provides pre-trained content moderation that detects harmful or inappropriate text categories, returning confidence scores for filtering user-generated content. It delivers the required accuracy without building custom models, unlike general-purpose services such as Bedrock or Rekognition.
A startup is using Amazon Bedrock to power a virtual assistant. They need to ensure that personally identifiable information (PII) is not included in the model's responses. Which feature should they enable?
Enable PII redaction in the Bedrock guardrails.
Bedrock guardrails apply PII redaction as a configurable filter that detects and masks sensitive data such as names, addresses and account numbers in both prompts and model responses. This directly satisfies the startup's requirement that personally identifiable information never appears in the virtual assistant's output, without retraining or altering the underlying foundation model.
Enable model invocation logging.
Configure a VPC endpoint.
Enable data encryption at rest.
Want more Applications of Foundation Models practice?
Practice this domain14% of exam · 6 sample questions below
A company uses Amazon SageMaker to train sensitive ML models. Which AWS service should they use to encrypt the training data and model artifacts at rest?
AWS Secrets Manager
AWS CloudHSM
AWS Key Management Service (KMS)
AWS KMS provides the customer-managed keys that encrypt SageMaker training data in S3 and model artefacts at rest, satisfying the at-rest encryption requirement. KMS integrates natively with SageMaker and S3, unlike services handling in-transit encryption or access control.
AWS Certificate Manager
A company uses Bedrock Guardrails to filter harmful content in a generative AI application. They need to prevent the model from discussing proprietary internal projects. Which Guardrail component should be configured?
Topic restrictions
Topic restrictions define denied topics using natural-language descriptions and example phrases, blocking discussion of named subjects such as proprietary internal projects. Content filters address harmful categories, not specific business topics, so topic restrictions satisfy this constraint.
Content filters
Grounding check
Word filters
A financial services firm needs to ensure that all calls to Amazon Bedrock APIs are logged for audit purposes. Which AWS service should they enable to capture API calls?
AWS CloudTrail
CloudTrail records API activity as management and data events, capturing every Bedrock API call with caller identity, timestamp and source IP. Enabling a trail delivers the audit logging the firm requires, satisfying the compliance constraint without altering application code.
Amazon S3 server access logs
Amazon CloudWatch Logs
AWS Config
A company wants to detect sensitive data such as PII in their training datasets stored in S3 before using them for model training. Which AWS service should they use?
Amazon Macie
Amazon Macie uses machine learning and pattern matching to automatically discover, classify and alert on sensitive data such as PII within S3 buckets, directly satisfying the requirement to scan training datasets before use. It continuously evaluates bucket contents against managed and custom data identifiers, providing the visibility needed prior to model training.
Amazon Inspector
AWS Shield
Amazon GuardDuty
A company uses SageMaker Clarify to detect bias in a deployed model. The monitoring must run automatically on a schedule. Which SageMaker feature should they use?
SageMaker Pipelines
SageMaker Experiments
SageMaker Data Wrangler
SageMaker Model Monitor
SageMaker Model Monitor runs scheduled bias and drift jobs against a deployed endpoint, integrating Clarify's bias metrics into automated monitoring schedules. This satisfies the stem's requirement for automatic, recurring bias detection, whereas Clarify alone provides analysis without native scheduling.
A company wants to use Amazon Bedrock to generate responses grounded in their proprietary knowledge base. They need to minimize hallucinations and ensure responses are based on the provided documents. Which feature should they enable?
Topic restrictions
Word filters
Grounding check
Grounding check measures whether each response is supported by the retrieved source passages, flagging or filtering ungrounded statements. This directly reduces hallucination by tying outputs to the proprietary knowledge base documents rather than model memory.
Content filters
Want more Security, Compliance, and Governance for AI Solutions practice?
Practice this domain20% of exam · 6 sample questions below
A data scientist wants to quickly build a supervised learning model for binary classification on a tabular dataset with 10,000 rows and 200 features. The dataset has some missing values and requires minimal code. Which AWS service should the data scientist use?
Amazon SageMaker Studio Lab
Amazon SageMaker Clarify
Amazon SageMaker Autopilot
SageMaker Autopilot automates algorithm selection, feature engineering and hyperparameter tuning for tabular classification, handling missing values and returning an explainable model with minimal code. It directly meets the binary classification requirement on the 10,000-row, 200-feature dataset.
Amazon SageMaker JumpStart
An ML team is deploying a real-time inference endpoint for a computer vision model using Amazon SageMaker. The model requires GPU acceleration for low latency. Which instance type should the team choose to minimize cost while meeting the GPU requirement?
ml.g5.xlarge
ml.g5.xlarge pairs an NVIDIA A10G GPU with four vCPUs, providing the required GPU acceleration at the lowest cost within the G5 family. Smaller GPU instances lack sufficient acceleration, while larger G5 sizes exceed the stated requirement.
ml.c5.xlarge
ml.p3.2xlarge
ml.p4d.24xlarge
A company needs to store large amounts of unstructured training data (images, videos) in a cost-effective manner while ensuring low-latency retrieval for training jobs running on Amazon SageMaker. Which storage solution should be used?
Amazon EFS
Amazon S3
Amazon S3 provides durable, low-cost object storage for unstructured images and video, and SageMaker training jobs read directly from S3 with high throughput and low latency. This satisfies both the cost-effectiveness and low-latency retrieval constraints in the stem.
Amazon RDS
Amazon EBS
An organization wants to detect anomalies in real-time streaming data from IoT devices. The data includes sensor readings, and the team plans to use a machine learning model. Which AWS service should be used to build and deploy the model with minimal operational overhead?
Amazon SageMaker
Amazon SageMaker provides fully managed infrastructure for building, training and deploying models, with built-in algorithms suited to streaming anomaly detection. It satisfies the stem's minimal operational overhead constraint by handling provisioning, scaling and endpoint hosting, letting the team focus on the model rather than servers.
AWS Glue
Amazon QuickSight
Amazon Kinesis Data Analytics
Which TWO services can be used to preprocess data for machine learning in AWS? (Choose two.)
AWS Glue
AWS Glue is a serverless ETL service that cleanses, transforms and catalogues data at scale using Spark jobs or Glue Studio, producing prepared datasets for ML training. It satisfies the preprocessing requirement by handling transformation and data quality tasks before model training.
Amazon Athena
Amazon SageMaker Data Wrangler
SageMaker Data Wrangler provides visual, low-code data preparation with over 300 built-in transforms, plus bias and anomaly analysis, exporting directly to SageMaker pipelines. It satisfies the preprocessing requirement by importing, transforming and featurising data before training.
Amazon Redshift
AWS Lambda
Which THREE statements about Amazon SageMaker Ground Truth are correct? (Choose three.)
It can only be used for text data.
It provides built-in workflows for image classification and object detection.
Ground Truth ships managed labelling interfaces and task templates for image classification and object detection, so teams avoid building custom annotation tooling. This built-in workflow support satisfies the stem's requirement that the statement describe a genuine Ground Truth capability.
It supports automated data labeling using active learning.
Ground Truth uses active learning: a model labels the confident examples automatically and routes only low-confidence items to human labellers, cutting cost. This automated labelling mechanism is a genuine Ground Truth feature, satisfying the stem's requirement.
It integrates with Amazon SageMaker to use the labeled data for training.
Ground Truth writes labelled datasets directly into Amazon S3 in augmented manifest format, which SageMaker training jobs consume natively. This tight integration lets the labelled output feed model training without custom transformation, satisfying the stem's requirement.
It can only use a public workforce from Amazon Mechanical Turk.
Want more Fundamentals of AI and ML practice?
Practice this domain24% of exam · 6 sample questions below
A company is building a chatbot using Amazon Bedrock and wants to ensure that the model generates responses consistent with its brand voice. Which technique should be used to provide the model with examples of desired responses without fine-tuning the model?
Fine-tune the model on a dataset of brand-compliant conversations.
Use prompt chaining to break down the conversation into multiple steps.
Implement a Retrieval Augmented Generation (RAG) system with brand documents.
Include few-shot examples in the system prompt to demonstrate the desired tone.
Few-shot examples in the system prompt steer Amazon Bedrock's model at inference time by conditioning it on demonstrations of the desired tone, satisfying the constraint of no fine-tuning. Unlike weight-updating approaches, this in-context prompting requires no training job, so brand-voice consistency is achieved immediately and cheaply.
A data scientist is using Amazon SageMaker to train a large language model from scratch. Which AWS service is most suitable for managing the training infrastructure, including automatic scaling and spot instance recovery?
AWS Lambda function.
Amazon SageMaker Notebook instance.
Amazon SageMaker Training job.
SageMaker Training jobs manage the underlying compute, handling automatic scaling and spot instance recovery natively. This satisfies the infrastructure-management requirement, unlike raw EC2 or manual cluster provisioning, which leave checkpointing and interruption handling to the data scientist.
Amazon EC2 with a custom setup.
A team is using Amazon Bedrock to generate images from text prompts. The generated images often contain artifacts and do not match the prompt description. Which combination of steps should the team take to improve image quality?
Fine-tune the model using SageMaker Ground Truth and increase the training epochs.
Increase the max token count and use a larger model variant.
Refine the prompt with more descriptive language and adjust the CFG scale and inference steps.
Prompt refinement supplies the missing descriptive detail that causes prompt mismatch, while the CFG scale controls how strictly generation adheres to the prompt and inference steps govern artefact reduction. Together these directly address both stated symptoms: artefacts and poor prompt alignment.
Use a different foundation model and increase the image resolution.
A developer is creating a generative AI application using Amazon Bedrock and needs to ensure that responses do not include toxic or harmful content. Which feature should be enabled?
Amazon CloudWatch Logs for prompt logging.
Amazon Virtual Private Cloud (VPC) for network isolation.
Amazon Bedrock Guardrails.
Guardrails applies configurable content filters and denied-topic policies to both prompts and responses, blocking toxic or harmful output at inference time. This satisfies the requirement that responses never include harmful content, without retraining or prompt engineering.
AWS Identity and Access Management (IAM) policies.
An organization is using Amazon Bedrock to power a customer service chatbot. They notice that the chatbot occasionally generates hallucinated information about product specifications. Which strategy should be implemented to reduce hallucinations?
Fine-tune the model on a dataset of product specification conversations.
Integrate a Retrieval Augmented Generation (RAG) system with the product catalog.
Integrating RAG grounds responses in retrieved product-catalogue content, so the model conditions on factual specifications rather than relying solely on parametric memory. This directly targets the hallucination source by supplying authoritative context at inference time, satisfying the requirement to reduce fabricated product details without retraining the foundation model.
Use more detailed prompts with explicit instructions to avoid speculation.
Increase the temperature parameter to make outputs more conservative.
A developer is using Amazon Bedrock's Claude model to summarize long documents. The developer notices that the summaries sometimes miss key points. Which parameter adjustment is most likely to improve summary completeness?
Increase the max_tokens parameter.
Truncation is the mechanism: if max_tokens is too low, the summary is cut off before covering all key points. Raising it allows the model to emit a complete summary, directly addressing the missed key points.
Increase the top_k parameter.
Increase the temperature parameter.
Increase the top_p parameter.
Want more Fundamentals of Generative AI practice?
Practice this domain14% of exam · 6 sample questions below
A financial services company uses Amazon Rekognition to verify customer identities. To ensure responsible AI practices, which measure should the company prioritize?
Use only black-box models to protect intellectual property
Increase model complexity to improve accuracy
Minimize the amount of training data collected
Regularly audit the model for demographic bias
Regular demographic bias audits directly satisfy responsible AI's fairness requirement by measuring whether Rekognition's identity verification accuracy differs across groups. This detects disparate error rates before they cause harm, which the other measures do not address.
A healthcare startup deploys a model to predict patient readmission risk using Amazon SageMaker. After deployment, the model shows higher false-positive rates for a specific age group. What is the most responsible first step?
Increase the prediction threshold for the affected group
Use Amazon SageMaker Clarify to detect bias in predictions
Amazon SageMaker Clarify quantifies bias across demographic groups using metrics such as disparate impact and equal opportunity difference, directly identifying the age-group disparity described. Detecting and measuring the bias before mitigation satisfies the stem's requirement for a responsible first step, since remediation cannot be targeted without first confirming which groups are affected.
Retrain the model with more data from the affected group
Immediately retire the model to prevent harm
A company uses an AI system to automate loan approvals. The model uses demographic features and achieves high accuracy, but the company wants to ensure compliance with responsible AI guidelines. Which practice best balances performance and fairness?
Use demographic features but with minimal monitoring
Use a complex black-box model and rely on post-hoc explanations
Remove sensitive attributes and monitor for proxy bias
Removing sensitive attributes directly addresses the fairness constraint, while monitoring for proxy bias catches indirect discrimination that demographic features create through correlated variables. This preserves predictive performance better than suppressing the model entirely, satisfying the stem's requirement to balance accuracy against responsible AI compliance.
Optimize the model solely for accuracy on historical data
A retail company uses a recommendation system that occasionally suggests inappropriate products to minors. Which responsible AI practice should be applied?
Implement human review of flagged recommendations
Human review of flagged recommendations directly satisfies the need to catch inappropriate outputs before minors see them. Automated filters alone cannot reliably judge context, so routing borderline cases to reviewers provides the oversight layer that mitigates harm, matching the responsible AI principle of accountability and safety in this retail scenario.
Rely solely on user feedback to improve
Disable the recommendation system entirely
Increase the volume of training data
A company uses Amazon Comprehend to analyze customer sentiment. They discover the model performs poorly on text with slang from underrepresented groups. What is the most responsible action?
Restrict model use to only standard English
Remove slang from input before inference
Adjust the confidence threshold only for those groups
Collect more representative training data including slang
Collecting representative training data that includes slang from underrepresented groups addresses the root cause: the model's vocabulary and patterns were learned from unrepresentative text. This improves sentiment accuracy for those groups rather than masking the disparity.
A bank uses an AI system to detect fraudulent transactions. The model has high precision but low recall for small transactions, potentially missing fraud. Which approach aligns with responsible AI?
Send all flagged transactions to customers for confirmation
Focus only on precision to minimize false positives
Tune the model to achieve an acceptable balance between recall and precision
Tuning to balance recall and precision directly addresses the low recall on small transactions, reducing missed fraud while keeping false positives acceptable. This aligns with responsible AI by mitigating harm from undetected fraud rather than optimising one metric alone.
Increase the detection threshold to reduce false positives
Want more Guidelines for Responsible AI practice?
Practice this domainThe AIF-C01 exam has 50 questions and must be completed in 90 minutes. The passing score is 700/1000.
Scenario-based questions covering exam objectives with detailed answer explanations.
The exam covers 5 domains: Applications of Foundation Models, Security, Compliance, and Governance for AI Solutions, Fundamentals of AI and ML, Fundamentals of Generative AI, Guidelines for Responsible AI. Questions are weighted by domain — higher-weight domains appear more on your actual exam.
No. These are original exam-style practice questions written against the official Amazon Web Services AIF-C01 exam objectives. They are not copied from the real exam. Courseiva focuses on genuine understanding, not memorisation of braindumps.
Courseiva tracks your accuracy per domain and routes you toward weak areas automatically. Free, no account required.