The work
You will help train and evaluate AI models by creating realistic patent prosecution scenarios and reviewing model responses. Your work will test whether outputs are legally sound, technically accurate, and strong enough to withstand USPTO review.
- Create scenarios involving prior art analysis, claim drafting, and office action responses.
- Write reference answers with clear legal and technical reasoning.
- Grade model outputs for claim quality, legal soundness, and technical accuracy.
- Identify drafting that looks plausible but would fail before the USPTO.
What it pays and takes
- Pay: $95 to $140 per hour, paid weekly through Stripe.
- Work type: Remote, fully asynchronous independent contractor role.
- Schedule: At least 10 hours per week, with no maximum and flexibility to adjust hours from week to week.
- Project: Ongoing.
- Location: Open to candidates in the countries specified for this role.
- Language: Clear written English is required.
- Required: Admission to a state bar and registration to practice before the USPTO.
- Required: At least two years of patent prosecution or patent litigation experience.
- Required: Strong claim drafting, office action response, and prior art analysis skills.
- Helpful: A technical degree in engineering or the sciences, experience across multiple technology areas, or experience with PTAB proceedings or patent litigation.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training is the human work behind systems that generate and review text. Contributors create examples, assess model responses, and explain what makes an answer accurate, useful, and legally sound; people with specialized expertise help models handle difficult professional work.