Polygon Segmentation Annotator (SAM Allowed)

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We hire and contract contributors to help teach AI systems by creating and validating the examples models learn from.

OpenTrain AI is the hiring and contracting organization for this role, and creating an OpenTrain account is free.

About AI training and why people do this work

AI training (also called data labeling or annotation) is the human work behind modern machine learning: people prepare accurate examples that teaching data-hungry models rely on. This field is a fast-growing way to work in tech.

Common benefits include fully remote, flexible schedules, and accessibility to entry-level contributors — many projects need no prior experience and let you choose your hours.

The role

You will create accurate polygon segmentations of a single target class per image. This is an entry-level contractor position that supports an ever-growing image dataset.

• Task type: Segmentation (single object per image)
• Annotation method: Polygons (you may use Segment Anything Model (SAM) but must correct its output manually)
• Tool: CVAT (our hosted instance)
• Dataset size: Approximately 18,000 images and growing
• Workflow: Continuous labeling as new data is added

What you'll do day to day

Perform polygon annotations for the specified target class in CVAT and verify your work. When using SAM or other assistive tools, check and correct every automated mask before submitting.

• Open images in CVAT and create or refine polygon masks for the single target object
• Use SAM optionally to generate initial masks, then manually correct boundaries and errors
• Follow provided labeling guidelines and quality checks before finalizing each label
• Continue labeling as new images are added to the dataset

Requirements and eligibility

This is an entry-level role suitable for contributors worldwide. No past experience or formal qualifications are required beyond the ability to work in CVAT and follow labeling instructions.

Employment type: Contractor; location: worldwide. There are no additional special requirements listed.

• Experience level: Entry level (no previous experience required)
• Languages: No language requirements specified
• Additional requirements: None listed

Compensation and payment

Payment is pay-per-label at a fixed rate per completed image annotation.

All payments are managed through OpenTrain according to the gig/contract terms when you join the project.

• Rate: $0.04 USD per labeled image (PAY_PER_LABEL)
• Work classification: Contractor

How work is organized and how to apply

Labeling is continuous: you can pick up tasks as new images arrive and complete them on your schedule within the project rules. Use CVAT for every annotation and apply the project’s polygon standard.

To apply, create an OpenTrain account (free), complete your contributor profile, and join the project from your dashboard. OpenTrain will provide access and the labeling guidelines after you are onboarded.

• Project access and instructions are provided through your OpenTrain account
• Expect to receive labeling guidelines and examples after joining