Formal Methods (Lean 4) Expert

About OpenTrain

OpenTrain is the #1 platform for building careers in AI training and data labeling. We help AI-training freelancers find specialized projects, consolidate opportunities, and build a unified portfolio so they can grow durable freelance careers in this fast‑moving field.

Why AI training matters

AI training (data labeling, annotation, and human feedback) is the human side of building modern AI systems. People prepare, curate, and evaluate examples that models learn from — everything from formal proofs to transcripts, translations, and image labels.

• 100% remote work you can do from anywhere with an internet connection.
• Flexible, part‑time friendly projects that fit around other commitments.
• Accessible entry points, with specialist roles paying more for deep expertise.
• Direct impact on how state‑of‑the‑art AI systems reason and behave.

The role

OpenTrain AI is hiring a Formal Methods (Lean 4) Expert to design and review challenging formal‑verification and theorem‑proving problems, then evaluate AI‑generated proofs and formalizations for correctness and rigor. This is a contractor, part‑time role (20+ hours/week) paid at $95/hour, open worldwide in English.

• Title: Formal Methods (Lean 4) Expert (Expert level)
• Work type: Contractor, Part‑time, 20+ hours/week
• Pay: $95/hour (PAY_PER_HOUR)
• Location: Fully remote; applicants worldwide (English required)

What you'll do

• Author challenging problems in theorem proving, program verification, and mathematical formalization.
• Review peer‑authored problems for clarity, difficulty, and ground‑truth correctness.
• Evaluate AI‑generated proofs, tactics, and formalizations and assign Accept / Revise / Reject verdicts.
• Provide detailed written rationales supporting every evaluation decision.

Requirements

• Deep background in formal verification or interactive theorem proving (expert level).
• Hands‑on experience with Lean 4 — ideally with mathlib contributions or equivalent work.
• Familiarity with other ITP systems such as Coq, Isabelle, or Agda is desirable.
• Strong understanding of type theory, mathematical logic, and program verification.
• Clear technical writing and meticulous attention to detail required for written review rationales.

Helpful background

Experience comparing model outputs for correctness and rigor will help you judge subtle differences in proofs and formalizations. Prior exposure to creating evaluation rubrics or writing reviewer feedback is a plus.

Compensation & logistics

This contract role pays $95 per hour and is expected to be 20+ hours per week. OpenTrain AI is the hiring and contracting organization for this role. Work is remote and available worldwide, with English required for reviews and written rationales. The primary data type is text; tasks include question/answering and evaluation/rating of model outputs.

• Payment type: PAY_PER_HOUR at $95/hour.
• Labeling focus: TEXT; label types include QUESTION_ANSWERING and EVALUATION_RATING.
• Employment: Contractor, Part‑time.

How to apply

Create a free OpenTrain account and submit your application for this role. When you apply, include any links or samples of Lean 4 work, mathlib contributions, or other formalization projects that demonstrate your experience. OpenTrain will manage hiring and contracting for the engagement.