The whole AI Practitioner track, sorted into the five scored domains. This is the foundational certification, so the track assumes no AWS AI background and defines its terms as it goes. Where a subject also appears in the professional-level Generative AI Developer track, the version here is the one pitched at this exam.
How to use this
Read each domain heading and its scope line, then run down the list. The order inside each domain is deliberate: every list starts at the deciding-level idea and works toward its refinements, so the post above you is the one the post below leans on. A line you can explain out loud, tick. A line that makes you hesitate is the next hour of revision. The quizzes and cards are the fastest way to close a gap.
The five scored domains and their weight:
| Domain | Weight |
|---|---|
| 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% |
| 5. Security, Compliance, and Governance for AI Solutions | 14% |
Weight is where the marks are, not where the difficulty is. Domain 3 is over a quarter of the score on its own.
Domain 1: Fundamentals of AI and ML (20%)
About 5 h 29 min of reading.
What machine learning is, what it is for, and how a model gets from data to production.
Anchor cheat sheet: ML Fundamentals and the SageMaker Suite · 41 min
What the words mean
About 1 h 29 min of reading.
- ☐ Telling AI, ML, Deep Learning, and Agentic AI Apart · 28 min
- ☐ What Your Data Decides Before You Pick a Model · 29 min
- ☐ Choosing How a Model Serves Its Predictions · 28 min
- ☐ Pop Quiz: Batch or Asynchronous · 1 min
- ☐ Pop Quiz: No Labels, No Target · 1 min
- ☐ Pop Quiz: Generative or Agentic · 2 min
Where AI earns its keep
About 1 h 39 min of reading.
- ☐ Turning a Business Question Into an ML Problem · 30 min
- ☐ Traditional Model or Foundation Model · 29 min
- ☐ When a Prediction Is the Wrong Answer · 37 min
- ☐ Pop Quiz: A Rule, Not a Prediction · 1 min
- ☐ Pop Quiz: Which Managed AI Service · 2 min
From data to a model in production
About 1 h 40 min of reading.
- ☐ Mapping an AI/ML Pipeline Onto AWS Services · 34 min
- ☐ How Much MLOps a Model Actually Needs · 28 min
- ☐ Measuring Whether a Model Earned Its Keep · 31 min
- ☐ Pop Quiz: Managed API or Self-Hosted · 2 min
- ☐ Pop Quiz: The Model Is Not Wrong, the World Moved · 2 min
- ☐ Pop Quiz: Ninety-Nine Percent and Blind · 3 min
Domain 2: Fundamentals of Generative AI (24%)
About 4 h 38 min of reading.
What a foundation model is, what it can and cannot do, and which AWS services build on one.
Anchor cheat sheet: Generative AI Foundations · 13 min
Generative AI groundwork
About 1 h 41 min of reading.
- ☐ Which Stage of the Foundation Model Lifecycle Is Yours · 30 min
- ☐ What Belongs in the Context Window · 28 min
- ☐ When an AI Application Becomes Agentic · 36 min
- ☐ Pop Quiz: What MCP Is For · 2 min
- ☐ Pop Quiz: Which Stage of the FM Lifecycle · 2 min
- ☐ Pop Quiz: Context Engineering or Fine-Tuning · 2 min
- ☐ Flash Card: What GenAI Gets Used For · 1 min
The AWS GenAI surface
About 1 h 08 min of reading.
- ☐ What a Token Costs and What Changes the Bill · 29 min
- ☐ The AWS Building Blocks for a GenAI Application · 34 min
- ☐ Pop Quiz: Input Tokens, Output Tokens, and the Bill · 2 min
- ☐ Pop Quiz: Bedrock, SageMaker AI, or JumpStart · 1 min
- ☐ Flash Card: Amazon Bedrock AgentCore · 1 min
- ☐ Flash Card: Strands Agents · 1 min
Fit, limits, and value
About 1 h 32 min of reading.
- ☐ Where Generative AI Helps and Where It Hurts · 28 min
- ☐ Proving a GenAI Feature Paid for Itself · 31 min
- ☐ What to Weigh When You Pick a Foundation Model · 29 min
- ☐ Pop Quiz: Naming the Generative AI Limitation · 2 min
- ☐ Pop Quiz: Which Metric Convinces Finance · 2 min
What generative AI is good for
About 4 min of reading.
- ☐ Pop Quiz: Which Job Suits a Foundation Model · 3 min
- ☐ Flash Card: Choosing a Foundation Model · 1 min
Domain 3: Applications of Foundation Models (28%)
About 5 h 37 min of reading.
Designing on top of a model: prompts, retrieval, customisation, and how you tell whether it works.
Anchor cheat sheet: AWS AI Services · 23 min
Retrieval and vector stores
About 34 min of reading.
- ☐ Answering Questions From Your Own Documents · 31 min
- ☐ Pop Quiz: Where the Embeddings Go · 2 min
- ☐ Flash Card: Vector Storage on AWS · 1 min
Training and customisation
About 1 h 12 min of reading.
- ☐ Deciding How Far to Customise a Foundation Model · 33 min
- ☐ Which Kind of Training a Foundation Model Needs · 35 min
- ☐ Pop Quiz: Facts That Change Weekly · 2 min
- ☐ Pop Quiz: Distillation or Continued Pre-Training · 2 min
Prompt engineering
About 1 h 08 min of reading.
- ☐ Picking a Prompting Technique for the Task · 36 min
- ☐ Keeping Prompts Safe and Versioned · 26 min
- ☐ Pop Quiz: Examples or Reasoning · 2 min
- ☐ Pop Quiz: Naming the Prompt Attack · 2 min
- ☐ Pop Quiz: Rolling Back a Prompt · 2 min
Designing on foundation models
About 1 h 13 min of reading.
- ☐ Choosing a Foundation Model for the Job · 36 min
- ☐ When an AI Agent Earns Its Place · 31 min
- ☐ Pop Quiz: One Prompt or an Agent · 3 min
- ☐ Pop Quiz: Capping the Response · 2 min
- ☐ Flash Card: Temperature, Top-P and Max Tokens · 1 min
Evaluating models and applications
About 1 h 07 min of reading.
- ☐ Judging Whether a Foundation Model Is Good Enough · 30 min
- ☐ Measuring Whether an AI Feature Is Working · 33 min
- ☐ Pop Quiz: ROUGE or BLEU · 2 min
- ☐ Pop Quiz: The Metric the Business Reads · 2 min
Domain 4: Guidelines for Responsible AI (14%)
About 4 h 49 min of reading.
Bias, fairness, transparency and explainability, and the tools AWS gives you for each.
Anchor cheat sheet: Responsible AI, Security, and Governance · 12 min
Data, bias, and variance
About 1 h 51 min of reading.
- ☐ Deciding Whether a Dataset Is Fit to Train On · 38 min
- ☐ Keeping Watch on Bias After Launch · 33 min
- ☐ Diagnosing a Model That Works for Most People · 36 min
- ☐ Pop Quiz: Imbalanced Data or Bad Labels · 1 min
- ☐ Pop Quiz: Overfitting, Underfitting, or Too Few Examples · 3 min
Transparency and explainability
About 1 h 25 min of reading.
- ☐ Explaining an AI Decision to the Person It Affects · 42 min
- ☐ When the Decision Has to Be Explainable · 33 min
- ☐ Pop Quiz: Interpretable or Explainable · 3 min
- ☐ Pop Quiz: Model Card, Service Card, or Artifact · 3 min
- ☐ Pop Quiz: The Thumbs-Down Button · 3 min
- ☐ Flash Card: AWS AI Service Cards · 1 min
Responsible AI features and tools
About 1 h 21 min of reading.
- ☐ Reducing the Legal Exposure of a Generative Feature · 39 min
- ☐ Picking a Model When Sustainability Is on the Scorecard · 36 min
- ☐ Pop Quiz: The Smaller Model and the Emissions Target · 2 min
- ☐ Pop Quiz: Who Owns the Generated Image · 3 min
- ☐ Flash Card: Amazon Bedrock Guardrails · 1 min
Domain 5: Security, Compliance, and Governance for AI Solutions (14%)
About 5 h 23 min of reading.
Securing the system, proving where data came from, and the evidence a regulator asks for.
Securing the AI system
About 1 h 22 min of reading.
- ☐ Matching an AI Security Worry to an AWS Control · 37 min
- ☐ Sorting AI Security Risks Into the Layer That Owns Them · 40 min
- ☐ Pop Quiz: What a VPC Endpoint Actually Does · 2 min
- ☐ Pop Quiz: Which Log Records What Was Said · 2 min
- ☐ Flash Card: AgentCore Identity and Policy · 1 min
Governance, compliance, and evidence
About 2 h 04 min of reading.
- ☐ Choosing the Service That Produces the Evidence · 35 min
- ☐ Deciding Where AI Data Lives and How Long It Stays · 42 min
- ☐ Setting Up Governance Before the First AI Feature · 41 min
- ☐ Pop Quiz: Config, CloudTrail, Inspector, or Artifact · 2 min
- ☐ Pop Quiz: Does Picking a Region Settle Residency · 2 min
- ☐ Flash Card: The Generative AI Security Scoping Matrix · 1 min
- ☐ Flash Card: What Bedrock Does With Your Prompts · 1 min
Grounding and output accuracy
About 35 min of reading.
- ☐ Keeping an Assistant From Making Things Up · 33 min
- ☐ Pop Quiz: When a Model Says It Is Confident · 2 min
Data provenance and secure data
About 1 h 22 min of reading.
- ☐ Four Things a Dataset Has to Be Before a Model Sees It · 40 min
- ☐ Showing Where an AI System’s Data Came From · 38 min
- ☐ Pop Quiz: Lineage, Catalogue, or Model Card · 2 min
- ☐ Pop Quiz: Access Control or Integrity · 2 min