AWS Certified AI Practitioner (AIF-C01)
exam objectives

99 practice questions in this domain

• 1.1Explain basic AI concepts and terminologies (AI, ML, deep learning, neural networks, computer vision, NLP, model, algorithm, training/inferencing, bias, fairness, fit, LLM, GenAI, agentic AI, inferencing types, data types, learning types)
• 1.2Identify practical use cases for AI (where AI/ML adds value, when it's inappropriate, selecting techniques like regression/classification/clustering, real-world applications, AWS managed AI/ML service capabilities, traditional ML vs. foundation models)
• 1.3Describe the AI/ML development lifecycle (AI/ML pipeline components, FM sources, production deployment methods, relevant AWS services per pipeline stage, MLOps concepts, model/business performance metrics)

119 practice questions in this domain

• 2.1Explain the basic concepts of generative AI (tokens, chunking, embeddings, vectors, prompt engineering, transformer-based LLMs, FMs, multi-modal/diffusion models, GenAI use cases, FM lifecycle, token-based pricing, context engineering, agentic AI concepts including MCP)
• 2.2Understand the capabilities and limitations of GenAI for solving business problems (advantages, disadvantages like hallucinations/nondeterminism, model selection factors, business value and metrics)
• 2.3Describe AWS infrastructure and technologies for building GenAI applications (Bedrock, SageMaker AI/JumpStart, Amazon Quick, Kiro, Strands Agents, Bedrock AgentCore; advantages of AWS GenAI services; infrastructure benefits; cost tradeoffs)

139 practice questions in this domain

• 3.1Describe design considerations for applications that use foundation models (FM selection criteria, inference parameters, RAG and vector database options, cost tradeoffs of customization approaches, role of AI agents)
• 3.2Choose effective prompt engineering techniques (concepts/constructs, techniques like chain-of-thought/zero-shot/few-shot, benefits and best practices, risks like prompt injection/jailbreaking, Bedrock Prompt Management)
• 3.3Describe the training and fine-tuning process for FMs (pre-training, fine-tuning methods, continuous pre-training, distillation, data preparation including RLHF)
• 3.4Describe methods to evaluate FM performance (human-in-the-loop, benchmark datasets, Bedrock Model Evaluation, metrics like ROUGE/BLEU/BERTScore/LLM-as-a-judge, business objective alignment)

69 practice questions in this domain

• 4.1Explain the development of AI systems that are responsible (bias, fairness, inclusivity, robustness, safety, veracity; Bedrock Guardrails; legal risks of GenAI; dataset characteristics; bias/variance effects; detection tools like SageMaker Clarify/Model Monitor/Amazon A2I)
• 4.2Recognize the importance of transparent and explainable models (transparent vs. non-transparent models, tools like SageMaker Model Cards/Clarify/Bedrock Model Evaluations, safety-vs-transparency tradeoffs, human-centered design for explainable AI)

69 practice questions in this domain

• 5.1Explain methods to secure AI systems (IAM, encryption, Macie, PrivateLink, shared responsibility model, Bedrock AgentCore Identity, Bedrock Guardrails; source citation/data lineage; secure data engineering; security/privacy considerations including prompt injection and data leakage; hallucination detection and grounding techniques)
• 5.2Recognize governance and compliance regulations for AI systems (AWS Config, Inspector, Audit Manager, Artifact, CloudTrail, Trusted Advisor; data governance strategies; governance protocols including the Generative AI Security Scoping Matrix)