The Work
You will prepare drone and maritime sensor imagery for AI and machine-learning systems. The work starts with still images and may expand to labeling individual frames from video.
You will work in a secure Navy environment, apply image-labeling guidelines consistently, and record unclear cases so questions can be addressed clearly.
- Review and correct machine-generated bounding boxes.
- Find and label objects that AI models miss.
- Apply labeling guidelines consistently across completed images.
- Deliver accurate annotated imagery while maintaining steady throughput.
- Record ambiguous cases and communicate guideline questions in writing and verbally.
- Adapt to video-frame annotation if the project expands.
What It Pays and Takes
This is a remote, part-time contractor role for contributors based in the United States. Prior AI experience is not required, and the role is suitable for entry-level contributors with relevant imagery knowledge and careful work habits.
- Pay: $30 to $60 per hour.
- Location: Remote, open to candidates based in the United States.
- Language: Fluent English communication in writing and speech.
- Security: Ability to obtain a DoD Common Access Card or participate in the sponsorship process.
- Required: Strong attention to detail and consistency during repetitive, accuracy-critical work.
- Required: Familiarity with aerial, maritime, or sensor-based imagery, including daylight or thermal formats.
- Expected: Comfort using dedicated labeling applications and learning new tools.
- Helpful: Experience with data labeling, bounding-box annotation, high-volume quality-controlled annotation, robotics, autonomous systems, or defense projects.