Text Classification Annotator

Your main job is to turn unstructured data into labeled training examples. Depending on the project, that might mean highlighting spans of text, drawing bounding boxes, assigning categorical labels, or ranking items by relevance. We provide annotation guidelines for every project, and you will go through a short calibration set before starting live work. The labels you produce feed directly into model training runs, so accuracy and consistency have a real impact.

Strong reading comprehension is non-negotiable. You should be comfortable reading dense material -- legal text, scientific abstracts, user reviews -- and making quick, accurate decisions about what category or tag applies. Prior experience with tools like Label Studio, Prodigy, or Scale AI is a plus but not required; our platform is browser-based and we will train you on it during onboarding.

Most projects run on weekly cycles. You will receive a batch of tasks on Monday, and the team expects completed work by Friday. Weekly calibration meetings happen over Zoom to align on edge cases and guideline updates. Minimum commitment is 15 hours per week; top contributors regularly log 30+.