Remote Software Expert; Scientific and Analytical - AI Trainer

Location

Remote (must have access to a physical Mac)

English

Mercor is supporting a high-priority data collection initiative aimed at improving how AI systems understand complex software interfaces and real-world, multi-step workflows. Current datasets lack the fidelity and expert grounding needed to reflect authentic professional software usage. This project addresses that gap by collecting high-quality screen annotations and screen recordings performed by experienced domain experts working in real digital environments.

• Record screen sessions demonstrating specific tasks, accompanied by clear verbal narration explaining each step
• Annotate screenshots of professional software by drawing precise bounding boxes around relevant UI elements
• Follow provided staging instructions to set up specific UI states prior to recording
• Use a custom capture tool to record workflows accurately and consistently
• Adhere closely to task guidelines to ensure data quality and usability

• Strong familiarity with professional software tools used in your domain, including:
  • MATLAB
  • Origin
  • Stata
  • EViews
• Detail-oriented and capable of following precise instructions
• Comfortable working independently and meeting tight deadlines
• Access to a physical Mac and can create a fresh macOS user profile if required

• Prior experience with data collection, annotation, or QA work
• Experience recording or documenting workflows
• Comfort working with new tools and staged environments

• Screen annotations are precise, consistent, and aligned with guidelines
• Screen recordings accurately capture realistic, expert workflows
• Tasks are completed efficiently while maintaining high quality
• Collected data is usable at scale for downstream AI research and development

We consider all qualified applicants without regard to legally protected characteristics and provide reasonable accommodations upon request.