Academy / AI Foundations

Path 1 · Course 2 · For everyone

How AI and Large Language Models Work

A clear, non-mathematical tour of machine learning, neural networks, transformers, training data and tokens — and why large language models sometimes get things confidently wrong.

📘 6 modules⏱ About 3 hours✅ Knowledge check per module🎓 Final assessment & certificate🎯 No prerequisites

0 of 6 modules complete

Start Module 1 Final assessment

Modules

  1. 1

    From rules to learning

    Traditional programming vs machine learning; data, features, models, training and inference; why AI took off.

    20 min
  2. 2

    Types of machine learning

    Supervised, unsupervised, reinforcement and self-supervised learning; testing, overfitting and measuring performance.

    25 min
  3. 3

    Neural networks and deep learning

    Artificial neurons, layers, learned features, parameters and scale, interpretability, and common architectures.

    25 min
  4. 4

    Language as numbers

    Tokens, embeddings, next-token prediction, temperature and context windows — with an optional tiny Python example.

    25 min
  5. 5

    Transformers and how LLMs are trained

    Attention, pre-training, instruction tuning, RLHF and Constitutional AI, reasoning models, multimodality, tools and agents.

    30 min
  6. 6

    Why LLMs get things wrong

    Hallucination, knowledge cut-offs, sycophancy, reasoning slips and bias — and how RAG, tools and review help.

    25 min