Do practical data work from wherever you are.
Remote, flexible, project-based work for people who are good at careful, repeatable tasks.
Annotation, data preparation, evaluation, research, verification and quality control — judged on precision rather than on your CV. Most projects can be done remotely; the schedule depends on the specific opportunity.
- Remote
- Flexible schedule
- No degree required
- No fees, ever
Much of the work behind modern AI is not glamorous.
Someone has to read the material. Someone has to label it. Someone has to compare outputs. Someone has to find the error. Someone has to verify whether an answer is actually correct. That work requires attention, consistency and judgment.
It is not clicking through a queue as fast as possible. Done badly it quietly ruins everything downstream, which is why precision is valued far above speed.
Typical work can include:
- annotation and labeling;
- data preparation;
- evaluation;
- verification;
- research;
- quality control;
- comparison of outputs;
- restructuring and cleaning information.
Some of what data work involves.
Labelling and annotation
Marking up text, images, audio or video according to a specification, and flagging the cases the specification did not anticipate.
Evaluation and comparison
Judging outputs against criteria: which is more accurate, more complete, more appropriate, and why.
Research and verification
Finding sources, checking whether a statement holds up, and recording what you found rather than what you assumed.
Structuring and cleaning
Turning inconsistent material into a form that can actually be used, without silently discarding the awkward parts.
Quality control
Reviewing other people's work, measuring agreement, and pushing back when a guideline produces nonsense.
Language work
Transcription, translation and localisation — where being genuinely fluent matters more than being fast.
You may be a fit if you are:
- a careful reader or writer;
- a researcher;
- a translator;
- an editor;
- someone already doing annotation or evaluation;
- someone changing careers;
- someone who prefers precise work to sales work.
No degree is required. These are the things that actually predict good work:
- you read carefully and notice when something does not fit;
- you follow a written guideline exactly, and say so when it is wrong;
- you are consistent across the hundredth item, not just the first ten;
- you would rather ask than guess;
- you can work independently, remotely, without supervision.
If you already do annotation or evaluation elsewhere, say so — it shortens qualification considerably.
There are no open task queues today and no published rates. Applying places you in the contributor pool; you may be considered for relevant opportunities as they appear. Applying does not guarantee work.
Rates are set by the specific opportunity and stated before work begins, together with the scope and the volume. We will not ask you to work first and discuss terms later.
We never ask applicants to pay for training, equipment, tests or access. If anyone claiming to represent us does, it is not us.
Apply as Data Contributor.
About 5 minutes · No CV upload · No fee. Say plainly what you have done — we are not looking for polish.