Fundamentals of AI and ML
AWS Certified AI Practitioner exam topic, about 20% of the exam. Updated .
The CEO Brief: play by chapter
Chapter 1: Speaking the same language
2 questions
Chapter 2: Not every problem needs AI
2 questions
Chapter 3: Is it working?
1 question
Storyline: The CEO Brief
A wholesale distributor's leadership team decides where AI earns its keep, where it does not, and how to tell if it is working.
The company is SMB Inc., a 120-person wholesale distributor. You meet Dana (CEO, no coding background), Amy (CFO, wants exact numbers and a clear return), Raj (data analyst in finance), Marco (operations manager, keen on AI) and Priya (head of customer success). Across three chapters, you are their advisor.
Chapter 1: Speaking the same language
AI is the broad goal of machines doing tasks that need human-like intelligence. Machine learning (ML) is the subset that learns patterns from data instead of following hand-written rules. Deep learning is ML built on multi-layer neural networks. Generative AI (GenAI) uses large models to create new content: text, images, code. Agentic AI goes a step further: it pursues a goal over several steps, decides what to do next, and takes actions through tools, often with a human approving the risky ones.
Inference is a trained model making predictions. Real-time inference answers single requests in milliseconds from an always-on endpoint. Batch inference scores large datasets on a schedule when nobody is waiting. Asynchronous inference queues large or slow individual requests. Serverless inference suits intermittent traffic and charges per request.
Chapter 2: Not every problem needs AI
AI predicts; it does not guarantee. When the rules are known and the answer must be exact, deterministic logic wins on cost, accuracy and auditability. When you do need a prediction, match the technique to the target: classification predicts a category, regression predicts a number, and clustering groups unlabelled data.
Chapter 3: Is it working?
Accuracy is the share of all predictions that are right, and it misleads when the event you care about is rare. Precision asks how many flagged cases were real. Recall asks how many real cases were caught. F1 balances the two. Pick the metric that matches the cost of each kind of mistake.