What is tasker?
The person who completes tasks on an AI training project, as opposed to the reviewer who checks them. Also called an attempter or contributor.
Tasker is platform shorthand for someone who does the tasks on an AI training project: writing prompts, rating responses, labeling data. Different platforms use different names for the same role, including attempter, contributor, annotator, rater and AI trainer.
Tasker and reviewer
Most projects run in two layers. Taskers produce the work; reviewers check a share of it against the rubric and give scores and feedback. Taskers with consistently high quality scores are often the ones invited to review, which can come with a higher rate.
What the job title tells you
When a listing says tasker, it usually means the work arrives as discrete tasks on a project, paid by the hour or by the task (see task-based pay). The amount of work depends on how many tasks are in the queue, which is why taskers talk about an empty queue.
Example
On a coding project, a tasker writes a prompt and grades two model answers. A reviewer later checks the grading and leaves a note that one bug was missed.
Looking for this kind of work? Browse open roles