The work
You will help improve AI systems by creating difficult mathematics problems and reviewing the answers they produce. Your work will focus on technical accuracy, proof quality, formalization, and practical mathematical reasoning.
You will work independently on project tasks and explain complex issues clearly in written feedback.
- Create advanced mathematics problems and scenarios based on proofs, formalization, and computational practice.
- Review AI-generated mathematical responses for accuracy, rigor, and alignment with professional practice.
- Find weaknesses in proofs and reasoning, including problems AI systems struggle to formalize.
- Write structured feedback that can guide improvements to AI model performance.
What it pays and takes
This is flexible, part-time contract work that can fit alongside another role. Tasks are completed remotely and at your own pace.
- Pay: Up to $47 per hour.
- Time: 20+ hours per week, with no set schedule.
- Location: Open to candidates in the United States.
- Language: Fluent written English.
- Experience: Advanced expertise in pure or applied mathematics and research experience shown through first-author publications.
- Skills: Ability to formalize proofs, explain complex mathematical ideas clearly, work independently, and maintain close attention to detail.
- Relevant depth: Formal methods, symbolic computation, optimization, numerical analysis, algebra, topology, number theory, or combinatorics.
- Helpful tools: Theorem provers such as Lean, Coq, or Isabelle, and symbolic tools such as SageMath or Mathematica.
- Consideration: Candidates who meet at least two of the three primary qualification areas may be considered.