Help improve frontier AI models by designing advanced mathematics problems, evaluating AI reasoning, developing computational solutions in Python, and contributing to formal theorem-proving tasks using Lean in a flexible remote contractor role.
Formal Methods (Lean 4) Expert
Job description
Role Overview
Mercor is seeking Formal Methods (Lean 4) Experts to help advance frontier AI systems by strengthening their formal reasoning capabilities. In this fully remote contract role, you will create expert-level theorem-proving and formal verification challenges, review technical content authored by other specialists, and evaluate AI-generated proofs and formalizations for correctness, rigor, and logical soundness.
This opportunity is ideal for researchers and engineers with deep expertise in interactive theorem proving, formal verification, and proof assistants such as Lean 4, Coq, Isabelle, or Agda.
Key Responsibilities
Develop Formal Methods Problems
- Design expert-level problems covering theorem proving, program verification, and mathematical formalization.
- Create technically rigorous evaluation tasks suitable for frontier AI models.
- Produce accurate reference solutions and formal specifications.
Review Technical Content
- Review problems created by other experts for clarity, correctness, and appropriate difficulty.
- Validate formal proofs and ensure reliable ground-truth solutions.
- Recommend improvements to increase technical quality and precision.
Evaluate AI Proofs
- Assess AI-generated proofs, tactics, and formalizations.
- Compare model outputs against expert standards.
- Deliver Accept, Revise, or Reject decisions with detailed written rationale.
Required Qualifications
- Strong background in formal verification or interactive theorem proving.
- Hands-on experience with one or more of:
- Lean 4
- mathlib
- Coq
- Isabelle
- Agda
- Familiarity with:
- Type Theory
- Mathematical Logic
- Program Verification
- Strong technical writing skills.
- Excellent attention to detail.
Preferred Qualifications
- Experience formalizing mathematical proofs.
- Experience reviewing formal verification projects or research.
- Interest in AI model evaluation and automated reasoning.
Compensation
- $95 per hour
- Weekly payments via Stripe or Wise
Work Arrangement
- Fully Remote
- Independent Contractor
- Flexible schedule
- Project-based engagement with opportunities for extension based on project needs and performance
You will be redirected to the company's website to complete your application.