Job Description
The Role:
We are building an edge AI system that monitors intersection safety in real time. We are expanding our detection capabilities and need a detail-oriented annotator to help us grow and maintain the training dataset that powers our models.
Responsibilities:
You will annotate video clips of intersections using CVAT, our annotation platform. The majority of your work involves reviewing and correcting AI-generated pre-labels rather than drawing from scratch — you are the quality layer, not the production layer. Day-to-day tasks include reviewing auto-generated bounding boxes across object classes (pedestrians, vehicles, and others), applying tracking corrections when objects are occluded or conditions change, and flagging edge cases for engineer review.
Requirements:
You have prior experience with CVAT, specifically with object tracking and interpolation workflows. You are comfortable working with video data across varied lighting and weather conditions. You have a high tolerance for repetitive, detail-oriented work and take quality seriously. Experience with YOLO-format datasets or prior computer vision annotation work is a strong plus.