AWS Certified AI Practitioner (AIF-C01)
Validate foundational understanding of artificial intelligence (AI), machine learning (ML), and generative AI concepts, terminology, responsible AI governance, and Amazon Bedrock / SageMaker solutions.
Official Exam Specification & Quick Facts
Question Difficulty Breakdown
Recommended Study Guides & Exam Preparation Blueprints
Reinforce your knowledge with domain breakdowns, trap question analyses, and architectural cheat sheets:
Exam Guidelines & Simulation Rules
Frequently Asked Questions (FAQ)
The AIF-C01 exam consists of 65 multiple-choice questions within 85 minutes. Cost is $75 USD (introductory rate), with a passing score of 700/1000 (70%).
Domains include: (1) Fundamentals of AI and ML (~20%), (2) Fundamentals of Generative AI (~24%), (3) Applications of Foundation Models (~28%), (4) Guidelines for Responsible AI (~14%), and (5) Security, Compliance, and Governance for AI Solutions (~14%).
Amazon Bedrock is a fully managed service that offers choice of high-performing foundation models (Anthropic Claude, Meta Llama, Amazon Titan, AI21, Cohere) via a single unified API with built-in security, RAG, and guardrails.
RAG integrates external enterprise data sources with foundation models via Knowledge Bases for Amazon Bedrock, grounding model responses with accurate cited facts and preventing hallucinations.
Amazon Bedrock is a serverless, managed API service for consuming pre-trained foundation models. Amazon SageMaker is a comprehensive platform for data scientists to build, train, tune, and deploy custom ML models from scratch.
Guardrails for Bedrock evaluates user prompts and model responses against customized policies, blocking harmful topics, filtering profanity, masking PII, and neutralizing prompt injection attacks.
Supervised learning trains models on labeled input-output pairs (classification/regression). Unsupervised learning discovers hidden patterns in unlabeled data (clustering). Reinforcement learning trains agents through trial-and-error rewards and penalties.
Foundation models are massive neural networks trained on broad data. They can be adapted through prompt engineering (zero/few-shot), RAG (dynamic contextual retrieval), or fine-tuning (updating model weights on domain-specific datasets).
Responsible AI principles include fairness (mitigating demographic bias), explainability (model interpretability), privacy and security (protecting PII and training data), transparency, and veracity (groundedness).
Vector embeddings convert unstructured text into numerical vector coordinates representing semantic meaning, allowing vector databases (OpenSearch Serverless, Pinecone) to perform fast semantic similarity searches.
The AIF-C01 certification is valid for three years from the date earned.
AIF-C01 is designed for business analysts, product managers, marketing professionals, software engineers, and leadership teams seeking foundational fluency in Generative AI and ML solutions on AWS.