Guide · Prompting

How to learn prompt engineering

Prompt engineering is the practice of getting reliable output from a language model on purpose rather than by luck. It covers choosing the right model, giving it the context it needs, structuring the request so the answer can be checked, and testing whether a change made things better or only different.

Last reviewed Who this is for: Anyone whose work already runs through a model and who is tired of results that were excellent yesterday and useless today.

How long does it take to learn?

The patterns take an afternoon to read and a week to feel. Evaluation, which is the part that separates a practitioner from someone with a good prompt, takes a month of doing it on real work. The tricks are quick. The judgement is not.

The fastest path, in order

  1. 01

    Learn what the model is doing

    Tokens, context windows, temperature, and the fact that it predicts rather than retrieves. Every reliable prompting habit is downstream of understanding this, and every superstition comes from not understanding it.

  2. 02

    Choose the model before you write the prompt

    A reasoning model, a fast cheap model and a long-context model fail in different directions. Half of what people call a prompting problem is a model selection problem, and it costs nothing to test three.

  3. 03

    Give context, not adjectives

    Examples of the output you want beat any number of instructions to be professional or thorough. Two or three worked examples in the prompt will outperform a paragraph of description almost every time.

  4. 04

    Make the output checkable

    Ask for structure you can validate: fields, a schema, a list of a stated length. An answer you can machine-check is an answer you can trust at volume, and it is what turns a prompt into something usable in a product.

  5. 05

    Build a tiny evaluation set

    Twenty real inputs with the answer you would accept. Run every prompt change against them. Without this you are not improving anything, you are only changing it and remembering the good runs.

  6. 06

    Learn where prompting stops

    Some problems are retrieval problems, some need tools, some need a different model. Recognising the ceiling saves the week people spend rewording their way towards a wall.

Tools worth your time

ToolWhat it is for
Anthropic ConsoleWorkbench for testing prompts against Claude models side by side, with variables.
OpenAI PlaygroundSame idea for GPT models: parameters exposed, output compared directly.
Google AI StudioFree surface for testing Gemini models, generous on long-context experiments.
promptfooOpen-source evaluation: run a prompt across cases and models, see what regressed.
LangSmithTracing and evaluation for prompts running inside an application.
A plain spreadsheetTwenty inputs, expected outputs, one column per prompt version. Unglamorous and sufficient.

Prompt libraries and template packs are worth about one afternoon of reading and nothing after that. The patterns are public and few. What is not transferable is knowing which one applies to your problem, and that only comes from running your own cases.

Mistakes that cost people weeks

The track that teaches it

Why this is the fastest way to actually get there

The tricks are freely available and mostly true. What almost nothing teaches is the evaluation habit, which is the only thing that turns a lucky prompt into a reliable one, and it is the difference between reading about prompting and being trusted to do it.

  • Model selection is taught before prompt patterns, because it is the bigger lever and it is almost always taught second or not at all.
  • Every lesson ends with your own cases, so you leave with an evaluation set rather than a folder of screenshots.
  • Kept current by editing text. Prompt guidance ages faster than any other AI subject, and a recorded course cannot follow it.

Questions people ask

The job title is fading and the skill is not. Models got better at reading sloppy instructions, which raised the floor. Getting consistent, checkable output on work that matters is still a practice, and it is now expected of people rather than hired for separately.

Prompting is what you say. Context engineering is everything you put in front of the model: documents, tool results, prior turns, and what you deliberately leave out. As models got better at instructions, most of the remaining difficulty moved to context.

No, for everyday use. Yes, once you want to evaluate prompts at volume or put one inside a product, because that means running the same prompt across many cases and comparing results.

Learn on at least two. The habits transfer, and using only one leaves you unable to tell a model limitation from a prompting mistake, which is the most common way people get stuck.

Prompt engineering as a discipline: how models work, model selection, the patterns that survive contact with real work, structured and checkable output, and building the evaluation set that tells you whether a change helped.

Vincoria opens soon.

One email when the doors open. Nothing else.

Founding members hear first. No spam.

Used only to tell you when Vincoria opens. One click to leave, from any email. Privacy policy

Next stepStart Lesson 01 · freeDevelop with AI · 9 min