Path 01 / Learning syllabus
AI basics
Learn what models do, evaluate their answers and build a small document assistant with traceable evidence and controlled tool access.
Explore the syllabus- 01What AI can and cannot doAvailable guide
- 02Run your first model requestAvailable guide
- 03Prompts, tokens and contextComing soon
Your course project
A small document assistant
Build a small assistant over synthetic service documentation. Each lesson adds a capability or checks whether it works; individual exercises also stand alone.
Before you begin
Start with the foundations.
For technically curious beginners. No prior AI knowledge or model training experience is required.
Each exercise states what you need and how to check its result. Use practice data and a disposable lab.
The ordered path
What you will learn
Follow the order, or open the guide that answers your question. Lessons marked Coming soon are not available yet.
- 01Available guide · 6 min read
What AI can and cannot do
Distinguish models, chat applications and agents, and explain why a fluent answer can still be wrong.
What you will try
Classify a document assistant’s answer, retrieved evidence and proposed action.
- 02Available guide · 19 min read
Run your first model request
Send a prompt, inspect the response, and understand hosted versus local inference. Use synthetic data and keep credentials out of prompts.
What you will try
Run a small document question twice and save both responses; explain why wording may differ.
- 03Coming soon
Prompts, tokens and context
Control instructions and supplied information; recognize context and output limits.
What you will try
Change one instruction while keeping the reference document fixed.
- 04Available guide · 10 min read
Weights, training and inference
Explain what a model learns, what a checkpoint contains and why supplying a document usually does not retrain it.
What you will try
Identify which changes affect request context and which require a training update.
- 05Coming soon
Evaluate an answer
Check claims and compare results against a small, repeatable test set.
What you will try
Score the assistant on answerable, ambiguous and unsupported questions.
- 06Coming soon
Choose a model for a task
Compare quality, latency, cost, privacy and operational requirements.
What you will try
Compare two configurations using the same questions and acceptance criteria.
- 07Coming soon
Embeddings and semantic search
Retrieve relevant passages and recognize misleading matches.
What you will try
Inspect a relevant result and a similar passage that does not answer the question.
- 08Coming soon
Ground answers with RAG
Inspect retrieval, supporting evidence and citations. Treat instructions inside retrieved documents as untrusted content.
What you will try
Check a supported answer and an unsupported answer, then add a document containing a conflicting instruction.
- 09Coming soon
Tools, workflows and agents
Follow an action from model proposal through validation, permission and execution.
What you will try
Inspect a proposed tool call using a harmless stub; verify the result without giving production access.
- 10Coming soon
Build a small document assistant
Combine retrieval and evaluation, measure cost, and handle missing evidence with explicit access limits.
What you will try
Deliver the assistant with a fixed test set, source references and documented failure cases.
