Applied AI

Course map · Outline

The whole course, in order

15 seasons. Two technical seasons on how AI works in production, then a leadership season that applies them. 16 lessons are live so far; each one links to the lesson before and after, so you can join anywhere.

  1. S1

    How LLMs work

    Technical season · 13 of 13 lessons live

    What a large language model does, from tokens and attention to hallucination and reasoning.

    1. Season 1: How LLMs work 5 slides
    2. One token at a time 13 slides
    3. Tokens 15 slides
    4. Embeddings 15 slides
    5. Attention 13 slides
    6. The transformer 16 slides
    7. Context windows 16 slides
    8. Field note: lost in the middle 18 slides
    9. Sampling 17 slides
    10. Hallucination 22 slides
    11. Field note: reasoning versus standard 20 slides
    12. When is a reasoning model worth the cost? 17 slides
    13. Season 1 recap 15 slides
  2. S2

    Retrieval and RAG

    Technical season, now running · 3 of 17 lessons live

    Grounding models in your own data with embeddings, search, chunking, reranking, and retrieval-augmented generation (RAG).

    1. Season 2: Retrieval and RAG 23 slides
    2. Why retrieval exists 16 slides
    3. Your first RAG system 26 slides
    4. Measuring it Coming soon
    5. Chunking Coming soon
    6. Search: keyword, vector, hybrid Coming soon
    7. Rewriting the question Coming soon
    8. Reranking Coming soon
    9. Filters, permissions and fresh documents Coming soon
    10. Building the prompt Coming soon
    11. Grounding and citations Coming soon
    12. Field note: does a citation rule change the answer? Coming soon
    13. Field note: swapping the embedding model Coming soon
    14. Field note: long context or retrieval? Coming soon
    15. Tracing a wrong answer Coming soon
    16. When RAG is the wrong answer Coming soon
    17. Season 2 recap Coming soon
  3. S3

    AI strategy and use case selection

    Leadership season

    Choosing where AI earns its place, building the business case, and knowing when not to use it.

    Lessons are announced in the season intro.

  4. S4

    Adapting models

    Technical season

    Prompting, structured output, fine-tuning, quantisation, and choosing the right model for the task.

    Lessons are announced in the season intro.

  5. S5

    Agent fundamentals

    Technical season

    How agents plan, use tools, and remember, including common design patterns and the Model Context Protocol (MCP).

    Lessons are announced in the season intro.

  6. S6

    AI economics

    Leadership season

    Build versus buy, managed versus self-hosted, lock-in, and the unit economics of AI features.

    Lessons are announced in the season intro.

  7. S7

    Data for AI

    Technical season

    Ingestion, document parsing, structured data, and the data quality every AI system depends on.

    Lessons are announced in the season intro.

  8. S8

    Evaluation

    Technical season

    Proving AI systems work, with golden sets, model-based judging, trajectory evaluation, and regression gates.

    Lessons are announced in the season intro.

  9. S9

    Building AI teams and operating models

    Leadership season

    Team structures, hiring, skills, and how AI teams work with the rest of the organisation.

    Lessons are announced in the season intro.

  10. S10

    Agents in production: reliability and traceability

    Technical season

    Tracing, failure modes, retries, and keeping humans in the loop.

    Lessons are announced in the season intro.

  11. S11

    Agents in production: security

    Technical season

    Prompt injection, guardrails, tool permissions, and limiting the blast radius.

    Lessons are announced in the season intro.

  12. S12

    Delivery, adoption, and change

    Leadership season

    Moving from pilot to production, aligning stakeholders, and getting AI adopted.

    Lessons are announced in the season intro.

  13. S13

    Agents in production: performance and cost

    Technical season

    Latency budgets, caching, inference optimisation, and token economics.

    Lessons are announced in the season intro.

  14. S14

    Agents in production: resilience

    Technical season

    What happens when an agent gets it wrong, and how to recover.

    Lessons are announced in the season intro.

  15. S15

    Governance and the regulated EU enterprise

    Leadership season

    The EU AI Act, data sovereignty, audit trails, and responsible AI in regulated industries.

    Lessons are announced in the season intro.

The field moves quickly. Seasons and lessons may be added as the course progresses.