We rebuilt PromptSewa — and your prompts were reset

Free course

Prompt engineering, from first principles

Twenty short lessons that build one skill: writing instructions a model can follow. Read them in order — each lesson uses the vocabulary the one before it established. 20 lessons, about 111 minutes of reading, and a certificate when you finish a lesson.

  1. 1 What a prompt actually is A prompt is a specification, not a question. This lesson reframes the whole course around that one idea. 5 min read
  2. 2 The anatomy of a great prompt Context, task, constraints and output shape — the four blocks every reliable prompt is built from. 6 min read
  3. 3 Role, audience and context A role is shorthand for a set of decisions. Use it to constrain judgement, not to flatter the model. 5 min read
  4. 4 Breaking a big ask into steps One prompt, one job. How decomposition turns an unreliable request into a sequence you can check. 6 min read
  5. 5 Constraints, formats and refusal rules The negative space of a prompt decides quality: what to avoid, what to refuse, and what shape to return. 6 min read
  6. 6 Variables and reusable templates Turn a good prompt into a tool: name the inputs, bound them, and keep the instruction fixed. 5 min read
  7. 7 Few-shot examples that actually teach One good example beats a paragraph of description — if it shows the decision, not just the format. 6 min read
  8. 8 Reasoning, step by step — carefully When asking for intermediate steps helps, when it hurts, and how to keep the reasoning checkable. 6 min read
  9. 9 Iteration: the second prompt is the good one Treat every first output as a diagnosis. Read the defect, change one thing, and keep the diff. 5 min read
  10. 10 Evaluating a prompt before you trust it A handful of fixed cases and a pass mark beat a feeling. Build a small test set and use it. 6 min read
  11. 11 Extraction prompts: getting facts out cleanly Pull structured fields from messy text without inventing the gaps. 5 min read
  12. 12 Transformation prompts: rewriting with control Summarise, shorten, restyle — without losing the facts that make the text true. 5 min read
  13. 13 Classification and labelling prompts Fixed labels, a rule for uncertainty, and a reason you can audit. 5 min read
  14. 14 Generation with style control Getting prose that sounds like you, without inventing facts to fill the shape. 5 min read
  15. 15 Image prompts: subject, style, light, camera Image models read nouns and relationships. Order your description so the scene survives. 6 min read
  16. 16 Video prompts: motion, camera and duration A video prompt describes change over time. Static descriptions produce a slideshow. 6 min read
  17. 17 Prompts for code Name the file, the behaviour and the constraint. Then ask for the test. 6 min read
  18. 18 Prompt security and injection When a prompt reads untrusted text, that text can give orders. Treat it as data, not instruction. 6 min read
  19. 19 Versioning and documenting prompts A prompt without a history is a prompt you cannot improve. Treat it like code. 5 min read
  20. 20 Building a prompt library that compounds Stop rewriting the same prompt. A small, organised library pays back every week. 6 min read