Skip to main content
Practice5 min read

Practice prompt engineering

Pick a track that matches your target role, work the five tasks, spend your attempts on iteration rather than rewrites, and move up the readiness bands.

Before you start

  • An Expert Hire practice account.
  • A target role in mind, since the tracks are organised by role.
Steps

5 steps, in order.

1

Pick the track that matches your target role

Prompt Engineering opens on a grid of assessments, filterable by mode, skill and level. Each is tagged Technical or Non-technical and carries a difficulty. Start with the one closest to the job you want, then branch out once you have a baseline.

  • AI-Powered Development

    Track
    Prompt Engineering
    Type
    Technical
  • AI for Software Engineers

    Track
    Prompt Engineering
    Type
    Technical, advanced
  • AI for Frontend Development

    Track
    Prompt Engineering
    Type
    Technical
  • AI for Data Analysts

    Track
    AI Research and Analysis
    Type
    Technical
  • AI for Business Intelligence

    Track
    AI Research and Analysis
    Type
    Non-technical
  • AI for Product Managers

    Track
    AI Communication
    Type
    Non-technical
  • AI for Professional Communication

    Track
    Prompt Engineering
    Type
    Non-technical
  • AI for Marketing Professionals

    Track
    AI Communication
    Type
    Non-technical
BackHomePrompt Engineering
Assessments
Prompt Engineering
Assess your AI prompting and problem-solving skills through real-world tasks.
AssessmentsMy Results
AI-Powered Development
Prompt Engineering
Technical
Solve real-world technical tasks using AI. Debug code, write documentation, generate tests, and refactor prompts for precise outputs.
Intermediate5 attempts
Start
AI for Professional Communication
Prompt Engineering
Non-technical
Use AI to draft, refine, and structure professional content. Rewrite emails, create proposals, summarize documents, and generate business outputs.
Intermediate5 attempts
Start
AI for Software Engineers
Prompt Engineering
Technical
Tackle real engineering challenges using AI, from architecture decisions to code reviews, migration planning, and incident response.
Advanced5 attempts
Start

Role-based tracks, each tagged technical or non-technical with a difficulty level.

2

Read the tasks before you begin

Opening an assessment shows what you are in for: five tasks, each with five attempts. On the frontend track, for example, the tasks are accessibility, CSS debugging, performance, component design and testing. Knowing the shape up front stops you burning attempts on task one.

3

Understand what an attempt is

The rules are strict and worth reading before you start, because attempts are the scarce thing here.

  • Every submitted prompt counts as 1 attempt

    What it means for you
    Think before you submit
  • Editing and resubmitting counts as a new attempt

    What it means for you
    There is no free correction
  • You can use voice-to-text or paste prompts

    What it means for you
    Draft outside the box if that helps
  • Scoring considers prompt quality, iteration and output

    What it means for you
    The path matters, not just the answer
  • Because scoring looks at iteration, a lucky first-try answer scores differently from a reasoned sequence. Do not treat attempts purely as retries.
4

Be specific, then iterate rather than restart

The single biggest lift comes from naming the task, the audience and the format you want out. Vague in, vague out. When a result is close but wrong, refine the prompt you already have instead of writing a fresh one from scratch. Refinement is what the scoring is looking for, and it is also what the job actually looks like.

  • State the format explicitly. Asking for a table, a diff or five bullets gets you something usable far more often than asking for an explanation.
5

Track your band, not your score

Results land in My Results, and your standing is reported as a readiness band rather than a bare number. There are four: Beginner, Emerging, Job Ready and Advanced. Moving up a band is the goal, and it is also the number your placement cell sees at cohort level.

  • Beginner

    Where you are
    Getting output, but not reliably what you asked for
  • Emerging

    Where you are
    Specific prompts, still rewriting rather than refining
  • Job Ready

    Where you are
    Consistent output, iterating deliberately
  • Advanced

    Where you are
    Using AI on genuinely hard, open-ended work
What you get

At the end of this guide.

A role-matched assessment across five real-world tasks.

A score that reflects prompt quality, iteration and output, not just the final answer.

A readiness band from Beginner through to Advanced.

A results history you can point at when an employer asks how you use AI.

Common mistakes

What goes wrong, and the fix.

Submitting a first draft to see what happens.

That is one of five attempts gone. Draft it properly, then submit.

Rewriting the prompt from scratch when the output is close.

Refine what you have. Iteration is part of what is being scored.

Leaving the output format unstated.

Say what you want back, and in what shape. It is the cheapest improvement available.

Starting on the hardest track to look good.

Start where your target role is, get a baseline band, then widen.

FAQ

Questions people ask on this one.

Still stuck? Talk to us and we will walk through your setup.

Try this in your own workspace.

The 30-day trial includes 75 Hire Credits, which covers a resume batch and a first round of interviews. Nothing charged for 30 days.

30-day free trial with 75 Hire Credits. Card required, nothing charged for 30 days.