AWS Certified AI Practitioner (AIF-C01)

Exam practice questions

15 / 192 questions

Free preview · Answers and explanations includedPage 1 of 2

Fundamentals of AI and MLAIF-C01 practice

A team proposes predicting rare equipment failures, but it has no reliable failure outcomes and cannot establish whether its current records represent the operating environment. What is the most appropriate next step?

Select one answerAAssess and improve data suitability before committing to a predictive solution.BEvaluate only on the records used for model training.CChoose a deployment endpoint before establishing the reliability of the outcomes.DTrain a larger model on the same unverified outcome labels.

Fundamentals of GenAIAIF-C01 practice

An assistant supplies a polished answer with a report title and page number. A reviewer cannot find the cited report, and the claim is absent from the supplied evidence. Which limitation is illustrated?

Select one answerAThe model can generate plausible but unsupported information.BA longer response necessarily contains more reliable evidence.CNatural-language fluency proves that retrieval succeeded.DEvery citation generated by a model is an authenticated source record.

Guidelines for Responsible AIAIF-C01 practice

A complex model is difficult to inspect directly, but a separate method estimates which features influenced a prediction. How should the team describe that result?

Select one answerAA post-hoc explanation that helps interpret the prediction without making the entire model inherently transparent.BA complete proof of every internal computation and its real-world cause.CEvidence that the model needs no accuracy evaluation.DA permanent guarantee that all future predictions are understandable.

Security, Compliance, and Governance for AI SolutionsAIF-C01 practice

A security reviewer needs to investigate which identity called an Amazon Bedrock API operation, when it occurred, and related request metadata. Which service is relevant, with logging configured for that operation's event type?

Select one answerAAmazon Inspector.BAWS Config.CAWS CloudTrail.DAWS Artifact.

Applications of Foundation ModelsAIF-C01 practice

An employee assistant must answer from approved procedures that change weekly. The team wants to update the knowledge source without retraining model weights for every edit. Which approach fits?

Select one answerARetrieve relevant current passages and supply them to the model as context through RAG.BFine-tune the model once and rely on those weights for every later handbook edit.CCache the original answer and reuse it whenever a similar question appears.DIncrease the output limit while keeping the same outdated evidence.

Applications of Foundation ModelsAIF-C01 practice

A team wants production behavior to use the exact prompt that passed acceptance testing while allowing further edits in development. Which Prompt management approach fits?

Select one answerAReference the latest editable draft from production.BCreate and reference an appropriate immutable prompt version for the release.CCopy a prompt nickname into release notes without retaining its definition.DSelect a new untested version automatically on each deployment.

Applications of Foundation ModelsAIF-C01 practice

An agent tries to update a support ticket, but the tool returns an error. What should the workflow do before telling the user the ticket was updated?

Select one answerAConfirm a successful tool result or report the failure and follow the defined recovery path.BGenerate a plausible ticket number without checking the system.CTreat the intended action as proof that the update happened.DHide every tool error and always report success.

Security, Compliance, and Governance for AI SolutionsAIF-C01 practice

A team needs to inspect full prompts and responses for supported Bedrock runtime calls in an approved debugging workflow. Which feature directly addresses this content-logging requirement?

Select one answerACloudWatch aggregate invocation-count metrics alone.BAmazon Bedrock model invocation logging, configured with appropriate destination and access controls.CCloudTrail API metadata alone, without model-content logging.DAWS Config resource-history records alone.

Fundamentals of GenAIAIF-C01 practice

Two models meet a document-labeling quality threshold. One is substantially faster and less expensive under the expected load. No additional capability of the larger model is required. Which selection approach is appropriate?

Select one answerASelect the slower model solely because it produces longer explanations.BAlways select the model with the largest parameter count.CIgnore measured latency because both models are generative.DChoose the model that meets the quality threshold and better fits the operating constraints.

Fundamentals of AI and MLAIF-C01 practice

An operations team receives scanned application forms and needs to extract printed text, key-value pairs, and tables for downstream processing. Which AWS service directly addresses this document-analysis requirement?

Select one answerAAmazon Lex.BAmazon Transcribe.CAmazon Textract.DAmazon Polly.

Fundamentals of AI and MLAIF-C01 practice

A supported SageMaker AI model processes a 600 MB input file and may take 25 minutes to finish. The application can submit the work and retrieve the result later. Which serving option is designed for this request pattern?

Select one answerASageMaker Asynchronous Inference.BSageMaker real-time inference with a synchronous request.CSageMaker Serverless Inference with its standard request limits.DSageMaker batch transform scheduled once every night.

Applications of Foundation ModelsAIF-C01 practice

A team wants a model to map short issue descriptions to its own unusual category names. Which prompt technique shows the desired mapping directly?

Select one answerAAdd examples from a different labeling convention without explaining the difference.BFine-tune the model by merely placing examples in one inference request.CSupply category names alone without definitions or demonstrations.DInclude several representative descriptions paired with the correct category labels.

Guidelines for Responsible AIAIF-C01 practice

A speech dataset contains clear recordings from one small group, but the application will serve users with varied accents and recording conditions. What should the team improve?

Select one answerAIncrease duplication of the original clear recordings.BRepresentation of the intended users and realistic audio conditions in the dataset and evaluation.CEvaluate only the easiest recording conditions to reduce score variation.DUse a random split of the same narrow collection as the sole coverage check.

Fundamentals of GenAIAIF-C01 practice

A semantic search application converts support articles into numerical representations so that articles with related meanings can be compared. What does this conversion produce?

Select one answerATraining labels that prove every article is accurate.BIAM policies that grant access to the articles.CFinal natural-language answers to every possible query.DEmbeddings represented as vectors.

Fundamentals of GenAIAIF-C01 practice

A project manager needs a short draft that captures the main decisions from a lengthy meeting transcript. The draft must remain faithful to the transcript and will be reviewed before circulation. Which generative AI task is appropriate?

Select one answerATranslate the complete transcript into another language.BSummarization of the supplied transcript.CExtract named people and organizations without summarizing their decisions.DCluster transcripts by topic without producing a decision summary.

Complete your exam preparation

Continue with the full question bank

Practise free with answers, explanations, and Community discussions. Join Community to unlock the complete question bank and timed mock exams.

Join Community Choose access for your certification

AWS Certified AI Practitioner (AIF-C01)

Exam practice questions

15 / 192 questions

Free preview · Answers and explanations includedPage 1 of 2

Fundamentals of AI and MLAIF-C01 practice

A team proposes predicting rare equipment failures, but it has no reliable failure outcomes and cannot establish whether its current records represent the operating environment. What is the most appropriate next step?

Fundamentals of GenAIAIF-C01 practice

An assistant supplies a polished answer with a report title and page number. A reviewer cannot find the cited report, and the claim is absent from the supplied evidence. Which limitation is illustrated?

Guidelines for Responsible AIAIF-C01 practice

A complex model is difficult to inspect directly, but a separate method estimates which features influenced a prediction. How should the team describe that result?

Security, Compliance, and Governance for AI SolutionsAIF-C01 practice

A security reviewer needs to investigate which identity called an Amazon Bedrock API operation, when it occurred, and related request metadata. Which service is relevant, with logging configured for that operation's event type?

Select one answerAAmazon Inspector.BAWS Config.CAWS CloudTrail.DAWS Artifact.

Applications of Foundation ModelsAIF-C01 practice

An employee assistant must answer from approved procedures that change weekly. The team wants to update the knowledge source without retraining model weights for every edit. Which approach fits?

Applications of Foundation ModelsAIF-C01 practice

A team wants production behavior to use the exact prompt that passed acceptance testing while allowing further edits in development. Which Prompt management approach fits?

Applications of Foundation ModelsAIF-C01 practice

An agent tries to update a support ticket, but the tool returns an error. What should the workflow do before telling the user the ticket was updated?

Security, Compliance, and Governance for AI SolutionsAIF-C01 practice

A team needs to inspect full prompts and responses for supported Bedrock runtime calls in an approved debugging workflow. Which feature directly addresses this content-logging requirement?

Fundamentals of GenAIAIF-C01 practice

Two models meet a document-labeling quality threshold. One is substantially faster and less expensive under the expected load. No additional capability of the larger model is required. Which selection approach is appropriate?

Fundamentals of AI and MLAIF-C01 practice

An operations team receives scanned application forms and needs to extract printed text, key-value pairs, and tables for downstream processing. Which AWS service directly addresses this document-analysis requirement?

Select one answerAAmazon Lex.BAmazon Transcribe.CAmazon Textract.DAmazon Polly.

Fundamentals of AI and MLAIF-C01 practice

A supported SageMaker AI model processes a 600 MB input file and may take 25 minutes to finish. The application can submit the work and retrieve the result later. Which serving option is designed for this request pattern?

Applications of Foundation ModelsAIF-C01 practice

A team wants a model to map short issue descriptions to its own unusual category names. Which prompt technique shows the desired mapping directly?

Guidelines for Responsible AIAIF-C01 practice

A speech dataset contains clear recordings from one small group, but the application will serve users with varied accents and recording conditions. What should the team improve?

Fundamentals of GenAIAIF-C01 practice

A semantic search application converts support articles into numerical representations so that articles with related meanings can be compared. What does this conversion produce?

Fundamentals of GenAIAIF-C01 practice

A project manager needs a short draft that captures the main decisions from a lengthy meeting transcript. The draft must remain faithful to the transcript and will be reviewed before circulation. Which generative AI task is appropriate?

Complete your exam preparation

Continue with the full question bank

Practise free with answers, explanations, and Community discussions. Join Community to unlock the complete question bank and timed mock exams.

Join Community Choose access for your certification