Turn your product-analytics experience into realistic data-analysis problems, worked solutions and grading rubrics used to train and evaluate AI models. Remote and freelance. How it works: apply, a short screening call, a trial task on a real rubric, then agreements and onboarding. We place shortlisted experts with our clients (AI labs, AI data partners and staffing partners). When you apply, you agree that we may share your profile with them for this and similar roles.
Role
Remote · Contract
Closes 31 Oct 2026
What you will do
Write analysis problems on acquisition, activation, engagement, conversion and retention, with a clear population, time window and metric.
Build A/B-test scenarios: hypothesis, success metrics, guardrails, control groups, and how to read the result, including uncertainty and practical significance.
Write SQL and Python solutions for cohort, funnel, customer lifetime value, retention and uplift analyses, with data-quality and validation checks.
Document assumptions, calculations and conclusions, and separate correlation from causation with suitable methods.
Write grading rubrics for reasoning, calculations and interpretation, listing acceptable alternatives and common errors.
Revise tasks when reviewers find ambiguity or unsupported conclusions.
What we look for
A completed master's degree or PhD in statistics, mathematics, data science, economics, computer science or a related field.
4+ years of hands-on analytical work as a product analyst, product or growth data scientist, or similar (titles can differ if the analysis was yours).
Professional SQL, Python and statistics, with examples of analysis you personally ran, not just dashboards.
Experience showing how an experiment, model or investigation changed a product or business decision.
Clear writing: you check assumptions and make your reasoning easy for a reviewer to follow.
Depth in at least one of: experimentation, growth analytics, funnels and cohorts, retention, customer lifetime value, segmentation, pricing, uplift modelling.
Not currently in a paid AI-evaluation, RLHF or annotation engagement elsewhere (past, clearly ended work is fine).
Use only material you're allowed to share; never include confidential employer or client data.
An up-to-date resume and LinkedIn profile that show the problem, your contribution, methods, validation and the decision it led to. Both are required.