androidengineers.Book a session

LEARN WITH A CLEAR DIRECTION

AI Basics: How LLMs Work

A free starter course for complete beginners: what AI, a model, and an LLM are, then tokens, the context window, temperature, embeddings, grounding, and agents — explained simply with small examples.

BeginnerAbout 90 minutes4 modules15 available lessons

Your path, one skill at a time.

New to AI? Start here — no background needed. Fifteen short lessons begin with the absolute basics: what AI is, what a model is, what an LLM is, and how apps talk to models. Then you'll learn the ideas behind every AI product: tokens, the context window, temperature and max tokens, structured output, embeddings, why models make things up, and how tools turn models into agents. Each lesson takes about five minutes and includes a tiny optional example. Finish ready to choose the AI Engineer, Agentic AI, or Forward Deployed Engineer path.

Self-paced. Fifteen lessons of about five minutes each.

What you’ll learn

  • Explain in plain words what AI, a model, and an LLM are.
  • Describe how an app sends prompts to a model through an API, and why keys stay secret.
  • Explain what a language model does and why it can be confidently wrong.
  • Estimate tokens and plan what fits in a context window.
  • Choose sensible settings such as temperature and max output tokens.
  • Describe embeddings, grounding (RAG), and how tools and agents work.

THE CURRICULUM

Everything you’ll learn, in order.

15 lessons across 4 modules. Open a module to see its lessons.

01The very basics5 lessons · About 30 min

What AI is, what a model is, what an LLM is, how apps talk to models, and the different kinds of models you can use.

  • AI, ML and generative AI
  • Models and training
  • LLMs
  • Prompts and APIs
  • Open vs closed, cloud vs on-device
  1. LessonWhat is AI?5 min
  2. LessonWhat is a model?6 min
  3. LessonWhat is an LLM?5 min
  4. LessonPrompts, responses, and APIs6 min
  5. LessonOpen and closed models, cloud and on-device6 min

Ready to move on when You can explain in plain words what a model is, what LLM stands for, and why an API key must stay on a server.

02How a language model works3 lessons · About 20 min

What a model actually does, how it reads text as tokens, and how much it can see at once.

  • Next-token prediction
  • Tokens
  • Context window
  1. LessonWhat a language model actually does5 min
  2. LessonTokens: how a model reads text6 min
  3. LessonThe context window6 min

Ready to move on when You can explain why a model might not know yesterday's news and why a long chat can forget its beginning.

03Talking to a model3 lessons · About 20 min

Messages and system instructions, the settings that shape output, and how to get data your app can use.

  • Roles and system instructions
  • Temperature, top-p, max tokens
  • Structured output
  1. LessonMessages, roles, and system instructions6 min
  2. LessonParameters and settings: temperature, top-p, and max tokens8 min
  3. LessonStructured output: getting data, not prose6 min

Ready to move on when You can pick a temperature for a task and explain why model output must be validated.

04Making models useful4 lessons · About 20 min

Embeddings, grounding answers in real documents, tools and agents — and where to go next.

  • Embeddings
  • Hallucinations and RAG
  • Tools and agents
  1. LessonEmbeddings: meaning as numbers6 min
  2. LessonWhy models make things up, and how grounding helps7 min
  3. LessonTools and agents in five minutes6 min
  4. LessonWhere to go next3 min

Ready to move on when You can explain how RAG reduces made-up answers and who runs a tool when a model calls one.