📄️ Prompt engineering
Prompt engineering with large language models, from a naive request to a versioned library evaluated on a labelled test set. Duration: 4h, with a running case study, a 40-question exam and a verifiable certificate.
📄️ 1. Anatomy
Module 1 of the prompt engineering premium course: the four elements a language model actually reads, what it ignores, and why the naive version of our running prompt fails.
📄️ 2. System prompt
Module 2 of the prompt engineering premium course: separate system framing from user input, understand which one the model trusts more, and stabilise behaviour across calls.
📄️ 3. Zero / one / few-shot
Module 3 of the prompt engineering premium course: choose between zero, one and several examples, order them well, and pay the token cost knowingly.
📄️ 4. Chain of thought
Module 4 of the prompt engineering premium course: when explicit reasoning helps a language model, when it hurts, and how to decompose a task into smaller model calls.
📄️ 5. Structured JSON
Module 5 of the prompt engineering premium course: produce JSON that validates against a schema, handle optional fields and unknown values, and retry on failure.
📄️ 6. Style and tone
Module 6 of the prompt engineering premium course: replace vague adjectives with measurable constraints, control length in sentences, and keep tone consistent across calls.
📄️ 7. Language pitfalls
Module 7 of the prompt engineering premium course: write prompts in the same language as the answer, avoid literal translations, and manage register, punctuation and units for English users.
📄️ 8. Injection defence
Module 8 of the prompt engineering premium course: direct and indirect prompt injection, delimiters, least-privilege tools, and what no prompt can guarantee on its own.
📄️ 9. Evaluation
Module 9 of the prompt engineering premium course: build a labelled test set, score per field, compare prompt versions with A/B, and detect regressions after a model change.
📄️ 10. Prompt library
Module 10 of the prompt engineering premium course: turn a working prompt into a versioned template, document its context of use, review it like code, and govern it inside a team.
📄️ Recap and exam
Complete recap of the prompt engineering premium course: anatomy, system framing, examples, reasoning, structured output, style, language, injection, evaluation and library, then the 40-question exam.