AI Practitioner mindmap

AWS Certified AI Practitioner (AIF-C01)

Foundational knowledge of AI, ML and generative AI concepts and AWS services. Check the official exam guide for current domains and weights.

• Domain 1: Fundamentals of AI and ML (20%)
  • Core concepts: AI, ML, deep learning, supervised, unsupervised and reinforcement learning
  • ML lifecycle: Data collection, training, evaluation, deployment and monitoring
  • Use cases: When AI fits, and when a rule-based solution is better
• Domain 2: Fundamentals of Generative AI (24%)
  • Concepts: Tokens, embeddings, transformers, prompts and foundation models
  • Strengths and limits: Adaptability versus hallucinations, nondeterminism and cost
  • AWS services: Amazon Bedrock, SageMaker JumpStart, PartyRock, Amazon Q
• Domain 3: Applications of Foundation Models (28%)
  • Model selection: Cost, latency, modality, context size and customization
  • Prompt engineering: Zero-shot, few-shot and chain-of-thought prompting
  • RAG: Knowledge bases and vector stores
  • Customization: Fine-tuning and continued pre-training
  • Evaluation: ROUGE, BLEU, BERTScore, human evaluation and benchmarks
• Domain 4: Guidelines for Responsible AI (14%)
  • Fairness and bias: SageMaker Clarify for bias and explainability
  • Safety: Bedrock Guardrails
  • Transparency: Model cards, explainability and trade-offs
  • Human review: Amazon Augmented AI (A2I)
• Domain 5: Security, Compliance, and Governance for AI Solutions (14%)
  • Security: IAM, encryption, PrivateLink and shared responsibility
  • Governance: AWS Config, CloudTrail, Audit Manager, Inspector, Artifact and Trusted Advisor
  • Data governance: Lineage, retention, residency and quality