The work
You will help evaluate AI systems that answer questions about veterans health care. The work focuses on whether questions sound realistic and whether answers are accurate, complete, useful, and current.
You will write clear feedback asynchronously so reviews remain consistent across the project.
- Review sample VA health care questions and flag questions that do not sound realistic for veterans or family members.
- Create rubrics for correct and complete answers about eligibility, enrollment, priority groups, income and means testing, care coordination, community care, the MISSION Act, and access to care.
- Compare AI-generated answers with those rubrics and flag errors, missing information, and outdated guidance.
What it pays and takes
This is a remote, short-term contractor project open to candidates in the United States.
- Pay: $30-$50 per hour.
- Schedule: Part time, 20 or more hours per week.
- Location: Remote and open to candidates in the United States.
- Language: Fluent in English.
- Experience: At least one year working with VA health care enrollment, eligibility determination, or care coordination.
- Required background: Former VA enrollment, eligibility, care coordination, or VA social work experience.
- Knowledge: VA priority groups, income and means testing, and community care.
- Skills: Clear, precise writing and the ability to judge whether guidance is accurate and useful.
- Helpful experience: Veteran outreach, helping veterans apply for health care, creating guides or training materials, or supporting rural or non-English-speaking veterans and families.
How it works
You apply here in about two minutes and finish on the hiring platform. The next step is a short AI interview (technical roles include a coding challenge). Passing it is how you qualify for paid tasks, so take it soon after you are invited.
The hiring platform's application for this job asks for your resume, your LinkedIn profile, your phone number, your location and answers to 4 screening questions.
About AI training work
AI training is the human work behind systems that answer questions and generate content. People review model responses, write evaluation rules, and identify mistakes, while subject-matter experience helps assess whether the guidance is accurate and useful.