Prepare test data
Build examples that show whether a change to the model, prompt, or data helped.
AI quality · alongside school · remote
AI Native (studies welcome)
You work with AI every day and find your own way around it. You do not have to know how to program. What matters is that you keep taking a problem apart until it makes sense and do not settle for the first answer a model gives you.
What you will do
You will keep an eye on the AI. You will prepare data that shows whether its answers are correct and find the places where the model goes wrong.
Find bad output
Review results, find patterns in errors, and state exactly when and why they happen.
Verify fixes
Check that a change fixed the problem without breaking other cases.
What we expect
You finish what you start
You look up what you do not know and say early when you need help.
You ask why
Knowing that something works is not enough. You want to know its limits.
You write precisely
Neither a model nor a colleague should have to guess what you meant.
You can sit with detail
Finding one bad answer among a hundred good ones gets dull. There is no way around it.
What we offer
Real work from the first week
No practice tasks for the drawer. A product or client project will use the result.
Hours that fit your studies
We agree on a scope you can carry alongside school. You can work from anywhere.
Paid work
We agree the rate based on scope and what you already know. This is not an internship paid in experience.
An experienced technical team
Twenty years of building software and leading teams. You get context and feedback directly.
Interested?
Tell us what you tried to build with AI and what did not work. Your CV does not need to be long.
Write to jobs@headers.cz