Data Annotator — Computer Vision (AI Safety Software)
Labeling images and video from safety cameras, to train AI detection models.
What we're looking for
Attention to detail. You will label thousands of images against the same rule. Image 3,000 has to be labeled the same way as image 3.
2 years of image or video annotation. Hands-on work drawing bounding boxes in a real tool — Label Studio, CVAT, Roboflow or similar.
You follow the guideline. When a rule doesn't cover a case, you ask. You don't guess and move on.
You can read basic numbers. Counts and percentages. You notice when a dataset looks thin or unbalanced.
Explicit written English. The team is remote. Most questions get asked in writing.
Nice to have: labeling CCTV or industrial site footage, correcting labels the AI drew, or working with COCO, YOLO or Pascal VOC formats.
What you'll do
Label images and video to a written guideline
Check labels for wrong classes, missed objects and loose boxes
Correct labels the AI has already drawn
Report unclear rules and missing data to the AI team
Location: Fully remote
Please apply this job if only you can start immidiately.