The work
You will create expert evaluation data for legal AI systems. The work calls for practitioner-level judgment across areas such as corporate, partnership, international, controversy, and state and local tax.
- Write realistic tax scenarios, from transaction structuring questions to controversy positions.
- Prepare reference answers that cite and weigh the Internal Revenue Code, regulations, and other relevant authorities.
- Evaluate model responses for technical accuracy, authority support, and professional defensibility.
- Identify answers that sound confident but misapply the Code or overlook controlling authority.
What it pays and takes
This is a fully remote, asynchronous independent contractor role. Work can scale up or down from week to week, with ongoing project work rather than single-task batches.
- Pay: $120 to $200 per hour.
- Schedule: At least 10 hours per week, with no fixed maximum.
- Payment: Weekly payment through Stripe.
- Location: Open to candidates in the countries listed for this role.
- Language: Strong written English.
- Education: JD degree.
- Licensing: Active bar admission in a US jurisdiction.
- Experience: At least two years of tax practice in a law firm, in-house, or government setting.
- Expertise: Depth in a substantive tax area such as corporate, partnership, international, controversy, or state and local tax.
- Helpful background: Transaction structuring, IRS controversy, cross-border tax, state and local tax, an LLM in Taxation, or a CPA credential.
- Skill: Ability to turn professional tax judgment into clear, teachable reasoning.
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 learn from examples, including writing scenarios and reviewing model responses. People with specialized knowledge help make those examples accurate, useful, and grounded in real professional standards.