Why Study with PlanetCert?
The Latest Questions
Practice questions and exam topics aligned with the current exam objectives.
Detailed Explanations
Go beyond the answer. Master the material with comprehensive learning and professional explanations for every concept.
AI-Powered Insights
Personalized preparation guidance that adapts to your performance and identifies weak spots automatically.
Exam Information
Official specifications published by Amazon
Exam Format
Registration
Validity
Skills this practice test covers
- Choose classification, regression or clustering for a given business use case.
- Describe the stages of an ML pipeline and the AWS services used at each stage.
- Compare RAG, fine-tuning and continued pre-training by cost and effort.
- Pick a prompt technique, such as few-shot or chain-of-thought prompting, for a given task.
- Match a responsible AI need with its tool, such as Amazon SageMaker Clarify for bias detection.
- Protect an AI workload with IAM permissions, encryption and AWS CloudTrail logging.
Explore related topics
AIF-C01 exam content
5 domains · 14 topics · 225 of 235 questions sorted by domain
Fundamentals of AI and ML20%55 questions
This domain sets out the basic words: AI, ML, deep learning, training, inferencing, bias and fit. Candidates also decide when ML adds value, when it does not, and which AWS services serve each stage of an ML pipeline.
Skills tested
- Define AI, ML, deep learning, neural networks, computer vision and NLP.
- Compare batch inferencing with real-time inferencing.
- Choose regression, classification or clustering for a given use case.
- Recognize when an AI/ML solution is not appropriate for a problem.
- Name the AWS services for each pipeline stage, such as Amazon SageMaker Data Wrangler for data preparation.
Official topics
- Explain basic AI concepts and terminologies
- Identify practical use cases for AI
- Describe the ML development lifecycle
Fundamentals of Generative AI24%50 questions
This domain explains what generative AI is, what it does well and where it fails. Candidates know tokens, embeddings and foundation models, and pick AWS services and infrastructure to build generative AI applications.
Skills tested
- Explain tokens, chunking, embeddings and vectors, and how transformer-based LLMs use them.
- Describe the foundation model lifecycle, from data selection to deployment and feedback.
- Identify the limits of generative AI, such as hallucinations and nondeterminism.
- Choose AWS services for generative AI, such as Amazon Bedrock, Amazon SageMaker JumpStart and Amazon Q.
- Weigh the cost tradeoffs of AWS generative AI services against their business value.
Official topics
- Explain the basic concepts of generative AI
- Understand capabilities and limitations of generative AI
- Describe AWS infrastructure and technologies for building generative AI applications
Applications of Foundation Models28%64 questions
The largest domain is about building with foundation models. Candidates choose a pre-trained model, set inference parameters, design RAG with a vector database, write prompts, plan fine-tuning and evaluate the results.
Skills tested
- Select a pre-trained model by cost, modality, latency, model size and input/output length.
- Explain how inference parameters such as temperature change a model's responses.
- Describe RAG and the AWS services that store embeddings in vector databases.
- Apply zero-shot, few-shot and chain-of-thought prompting, and know the risks of prompt injection and jailbreaking.
- Evaluate a foundation model with metrics such as ROUGE and BLEU, and against business objectives.
Official topics
- Describe design considerations for applications that use foundation models
- Choose effective prompt engineering techniques
- Describe the training and fine-tuning process for foundation models
- Describe methods to evaluate foundation model performance
Guidelines for Responsible AI14%30 questions
This domain is about AI that is fair, safe and explainable. Candidates identify the features of responsible AI, detect bias in datasets and models, and weigh transparency against safety.
Skills tested
- Identify features of responsible AI, such as fairness, inclusivity, safety and veracity.
- Detect bias with Amazon SageMaker Clarify, and route predictions to people with Amazon Augmented AI (A2I).
- Explain the legal risks of generative AI, such as intellectual property claims and biased outputs.
- Tell transparent models from black-box models, and document a model with Amazon SageMaker Model Cards.
Official topics
- Explain the development of AI systems that are responsible
- Recognize the importance of transparent and explainable models
Security, Compliance, and Governance for AI Solutions14%26 questions
This domain protects AI systems and shows that they follow the rules. Candidates secure data and access, record where data came from, and use AWS services for governance and compliance.
Skills tested
- Secure an AI system with IAM roles, policies and permissions, encryption and AWS PrivateLink.
- Find sensitive data in Amazon S3 with Amazon Macie before it is used for training.
- Document data origins with source citation and data provenance tracking.
- Use AWS Config, AWS CloudTrail, AWS Audit Manager and AWS Artifact for governance and compliance.
- Describe data governance strategies such as retention, residency, logging and monitoring.
Official topics
- Explain methods to secure AI systems
- Recognize governance and compliance regulations for AI systems
Requirements
- No formal prerequisites: anyone can book the exam.
- Recommended: familiarity with AI/ML concepts and use cases, and up to six months of exposure to AI/ML on AWS.
- Recommended: a basic understanding of AWS Cloud fundamentals, for example from AWS Certified Cloud Practitioner.
About the AIF-C01 exam
AIF-C01 tests whether a person understands AI, machine learning and generative AI well enough to pick the right AWS service for a business problem. It is written for people who use or choose AI solutions, such as business analysts and product managers, more than for those who build models. The two largest domains are about generative AI and foundation models: RAG, prompt engineering, fine-tuning and evaluation. The other domains cover AI and ML basics, responsible AI, and security and governance. Passing earns the foundational AWS Certified AI Practitioner certification.
Applications of Foundation Models has the most questions in the PlanetCert practice test, with vector databases, prompt techniques and inference parameters such as temperature. Many explanations turn on one service name, for example why Amazon SageMaker Clarify suits a bias check and Amazon Augmented AI (A2I) suits human review. Study Mode suits the customization questions, where prompt engineering, RAG and fine-tuning each fit a different need. Good questions for Brainy AI Assistant include how a RAG design uses a vector database, or what a higher temperature does to a model's answers.
How to earn the certification
- Earns
- AWS Certified AI PractitionerFoundational
The certification is valid for three years.
- Before
- CLF-C02 AWS Certified Cloud Practitioner
Suggested, not required. It covers the AWS Cloud basics that many AIF-C01 questions take for granted.
- CLF-C02 AWS Certified Cloud Practitioner
- After
- MLA-C01 AWS Certified Machine Learning Engineer - Associate
The associate step for those who move from understanding ML to building and deploying models on AWS.
- DEA-C01 AWS Certified Data Engineer - Associate
The associate step for those who build the data pipelines that ML and generative AI depend on.
- SAA-C03 AWS Certified Solutions Architect - Associate
A broader associate step about designing AWS architectures, AI workloads included.
- DVA-C02 AWS Certified Developer - Associate
The associate step for developers who build and deploy applications on AWS.
- MLA-C01 AWS Certified Machine Learning Engineer - Associate
How to prepare for AIF-C01
- 1
Define the basic AI words first
The first domain defines AI, ML, deep learning, inferencing, bias and fit. Later domains build on these words, so each one should be clear in one plain sentence.
- 2
Try prompts in Amazon Bedrock
Send one prompt to a model in Amazon Bedrock with a low and a high temperature, and compare the answers. Then connect a small set of documents through Amazon Bedrock Knowledge Bases to see RAG at work.
- 3
Weigh RAG against fine-tuning
Applications of Foundation Models is the largest domain. Compare prompt engineering, RAG, fine-tuning and continued pre-training by cost, effort and the problem each one solves.
- 4
Know which tool does which job
Amazon SageMaker Clarify checks bias, Amazon SageMaker Model Cards document models, Amazon Augmented AI (A2I) adds human review and AWS Artifact supplies compliance reports. Drill these pairs in flashcard mode.
- 5
Run the full test under time
A timed run of the AIF-C01 practice test shows which service names are still unclear. For each one missed, ask Brainy AI Assistant to compare the service chosen with the right one.
Start a timed AIF-C01run
Who AIF-C01 is for
- Business analysts and product managers who help decide where AI fits in a product or process.
- Project managers and line-of-business managers who lead projects that use AI on AWS.
- Marketing and sales professionals who discuss AI services with customers.
- IT managers and support staff whose teams start to run AI workloads on AWS.
Typical job titles: Business analyst · Product manager · Project manager · IT manager · Marketing professional · IT support professional
AIF-C01 glossary
18 terms from the exam objectives and the vendor's official material, each defined in a sentence.
- Amazon Augmented AI (A2I)
- The AWS service that sends selected model predictions to people for review.
- Amazon Bedrock
- The AWS service that offers foundation models from several providers through one API, for building generative AI applications.
- Amazon Bedrock Agents
- The Amazon Bedrock feature that lets a model carry out multi-step tasks by calling APIs and data sources.
- Amazon Bedrock Knowledge Bases
- The Amazon Bedrock feature that connects a model to an organization's own documents, so that answers can draw on them.
- Amazon Macie
- The AWS service that finds sensitive data, such as personal information, stored in Amazon S3.
- Amazon SageMaker Clarify
- The SageMaker tool that measures bias in data and models and helps explain model predictions.
- Amazon SageMaker JumpStart
- A hub in Amazon SageMaker of pre-trained models and solution templates that can be deployed or fine-tuned.
- Amazon SageMaker Model Monitor
- The SageMaker tool that watches models in production and raises an alert when data or quality changes.
- AWS Artifact
- The AWS portal where customers download AWS's own compliance reports and agreements.
- AWS CloudTrail
- The AWS service that records the API calls made in an account, which gives an audit trail of who did what.
- BLEU score
- A metric that scores generated text by how closely its wording matches a reference text, often used for translation.
- Chain-of-thought prompting
- A prompt technique that asks the model to work through a problem step by step before it gives an answer.
- Chunking
- Splitting long documents into smaller pieces before they are turned into embeddings and stored for retrieval.
- Context window
- The amount of text, counted in tokens, that a model can take into account at one time.
- Continued pre-training
- Training a foundation model further on unlabeled data from one field, so that it learns that field's language.
- Data poisoning
- Tampering with training data on purpose, so that the model learns wrong or harmful behavior.
- Retrieval Augmented Generation (RAG)
- A design in which the application first finds relevant documents and then passes them to the model with the prompt.
- Vector Databases
- Databases that store embeddings and quickly find the ones closest in meaning to a query.