Computer Vision Data Labeler

This role is about precision marking. You will use our web-based annotation tool to identify and outline objects in images, classify their properties, and annotate spatial relationships. Some projects involve simple bounding boxes; others require polygon segmentation at near-pixel accuracy. The quality bar is high because the models trained on your annotations will make real-world decisions -- a mislabeled pedestrian in a driving data set is not a minor error.

We look for annotators who balance speed and accuracy well. Some people are very precise but too slow; others are fast but sloppy. The sweet spot is consistent, efficient annotation that stays within quality thresholds. A background in any visually detailed field -- illustration, photography, radiology, geographic mapping -- is a strong signal.

Projects run in weekly sprints with clear volume targets. You will know at the start of each week how many images or frames to annotate. Most annotators work 15-25 hours per week. A brief weekly call with the project lead covers guideline updates and quality feedback.