Computer Vision QA Specialist

What you'll do

You are the second pair of eyes on computer vision labels. Annotators draw bounding boxes, segmentation masks, polygons, keypoints and video tracks; you check them for missed objects, wrong classes, loose boundaries, duplicates and inconsistent attributes, then fix them or write up what is wrong. The guidelines are detailed, with rules for occlusion, truncation, blur, small objects and overlaps, and you escalate ambiguous cases rather than guessing. Work comes in large batches, so consistency matters as much as the single hard case.

Who fits

• Experience reviewing or creating image or video annotations is preferred, not required
• A sharp eye for missing, mislabeled or poorly placed annotations
• Comfort with edge-case rules and detailed guidelines
• Working English to read guidelines and write feedback
• Ability to commit steady hours; platform experience (SuperAnnotate, Labelbox, CVAT, Scale, Appen, Toloka) is preferred

Open to Bangladesh and India only.

What it pays

The listing states up to $2.5 an hour, a figure we have taken as written. It is far below every other role in this group, so decide whether the rate works for you before you invest in the interview. There is no floor stated. The listing does not publish payment method or cadence; SuperAnnotate's general terms are on the company page.

Worth knowing

Good:

• Low barrier: no degree or years required
• Skills transfer to other annotation QA work
• Remote and hourly

Less good:

• The stated ceiling is very low, and "up to" means you may earn less
• Large batches reward speed, which can conflict with careful review
• Hourly work on SME projects is tracked with monitoring software, per the company page.

About this listing

Read from the SME Careers jobs board on 27 October 2026; the board shows no posting date.

See SuperAnnotate.