The Work
You will turn professional legal and compliance judgment into training data for AI systems. The work includes writing challenging scenarios, defining what strong answers must contain, and reviewing model responses for sound reasoning and appropriate risk analysis.
- Write realistic legal and compliance questions that require expert analysis.
- Design rubrics covering key issues, standards, reasoning, completeness, exceptions, jurisdiction, risk, and critical errors.
- Evaluate AI responses for accuracy, relevance, completeness, reasoning quality, and professional rigor.
- Identify unsupported conclusions, incorrect rule application, missing exceptions, faulty assumptions, and poor risk prioritization.
- Review peer rubrics for ambiguity, redundancy, hidden assumptions, excessive subjectivity, and inaccurate legal expectations.
- Create reference analyses, answer frameworks, critiques, preference data, and reference-answer pairs for RL and SFT workflows.
- Distinguish clearly between factual analysis, legal interpretation, professional judgment, and uncertainty.
What It Pays And Takes
This is a US-based part-time contract role. The work is listed as entry level, while the qualifications call for substantial legal or compliance knowledge and professional experience.
- Pay: $120 per hour.
- Time: 20+ hours per week.
- Location: Open to candidates in the United States.
- Work type: Contractor and part time.
- Education: JD, LLB, LLM, law degree, compliance qualification, or equivalent professional background.
- Experience: At least three years in law, regulatory compliance, legal research, policy, risk, privacy, financial compliance, contracts, or a related field.
- Language: Professional English proficiency at minimum C1 level.
- Expertise: A clearly defined legal, regulatory, or compliance domain.
- Skills: Assess controlling issues, rule application, material risks, legitimate alternative interpretations, terminology, authority, jurisdiction, factual assumptions, materiality, and conclusions.
- Rubric writing: Define required issues, standards, exceptions, omissions, and reasoning quality.
Helpful Background
Experience in legal education, examination writing, peer review, legal editing, quality assurance, policy drafting, compliance testing, supervisory review, AI evaluation, legal AI, data annotation, RLHF, model benchmarking, or rubric-based scoring can support this work.
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 and rating responses, evaluating reasoning, and preparing structured feedback. People with legal and compliance expertise are needed because they can judge rules, exceptions, risks, and uncertainty that general reviewers may miss.