Product Manager Expert for AI Evaluation

You will create expert evaluation data for AI models handling product management tasks. The work tests whether a model can understand users, make sound priorities, explain tradeoffs, and choose meaningful actions instead of producing polished but shallow plans.

• Write realistic product scenarios that move from an unclear problem to a shipped decision.
• Create reference answers that explain what to prioritize, cut, pursue, and why.
• Rate model responses about product specifications, roadmaps, and strategy.
• Identify plans that miss important tradeoffs, overlook users, or lack meaningful reasoning.
• Apply practical product management judgment to AI training and evaluation data.

What it pays and takes

This is remote, fully asynchronous independent contractor work that can continue on an ongoing basis. Pay is issued weekly through Stripe.

• Pay: $60 to $110 per hour.
• Time: At least 10 hours per week, with no maximum.
• Location: Open to candidates in the countries listed for this role.
• Language: Fluent written English is required.
• Experience: At least two years owning a live product or feature area as a product manager.
• Required background: Shipped product work, including specifications, launches, and outcomes you can explain.
• Skills: Strong knowledge of roadmaps, PRDs, analytics, experiments, and product team working practices.
• Helpful background: B2B SaaS, consumer, or platform products; technical work with engineering teams; or AI and data products.

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 model responses. Experienced professionals are paid to provide the judgment and subject knowledge needed to evaluate specialized AI behavior.