You will create realistic litigation work product for multi-turn commercial real estate disputes. The work helps train and evaluate AI systems, and prior AI experience is not required.
You will work under senior direction, revise drafts after feedback, and keep the case theory consistent across multiple rounds.
- Draft motions, briefs, demand letters, and discovery responses.
- Prepare turn-by-turn deliverables that reflect drafts submitted for partner review.
- Create rubric items with your drafts to explain what makes a strong response to opposing counsel.
- Apply senior feedback efficiently during each round.
- Flag procedural uncertainties, missing information, and unclear facts that need resolution.
What it pays and takes
This is a part-time contractor role open to candidates in the United States. The listed experience level is entry level, while the role requires substantial litigation drafting experience.
- Pay: $100 per hour.
- Time: Less than 20 hours per week.
- Location: Open to candidates in the United States.
- Language: Professional fluency in English, with strong legal writing and analysis.
- Bar status: An active or lapsed U.S. bar license in at least one jurisdiction.
- Experience: At least three years of post-JD litigation experience involving real drafting, discovery, and motion practice.
- Subject knowledge: Commercial real estate litigation involving leases, transactions, development, landlord-tenant matters, zoning, or financing.
- Writing ability: Able to draft complex litigation documents under direction.
- Helpful background: Firm experience in BigLaw, a boutique, or a mid-market firm; prior AI-based legal projects or federal court commercial 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 evaluate text. Legal professionals contribute by writing realistic examples, reviewing model responses, and setting clear standards for accurate, useful output.