What the work actually involves

You are given documents that sit inside the operations and program-management world — a weekly program status report, a project charter, a risk register, an SOP for a fulfilment process, a resourcing model in Excel, a steering-committee deck — and asked to judge them. Sometimes the artifact was written by a model and you are scoring it; sometimes you are comparing two versions and choosing which a real PMO would accept; sometimes you are writing the reference version yourself so a model has something correct to learn from. The output is rarely a number alone. It is a rubric score plus written reasoning that names the specific defect: the milestone dependency that cannot be sequenced that way, the RAG status marked green while the mitigation owner is blank, the capacity plan that quietly assumes 100% utilisation.

The distinguishing skill is not writing project documents — plenty of people can. It is articulating why a plausible-looking document would fail in practice, in language another reviewer can audit. Vague feedback like "lacks detail" is worthless to the training pipeline. "The escalation path names a role that has no decision authority over vendor spend, so this risk would sit unresolved for a full sprint" is the standard.

What the platform screens for

  • Verifiable operating history. Which programs, what scale, what your actual accountability was — budget, headcount, cross-functional scope. "Supported the PMO" invites follow-up you should be ready for.
  • Artifact fluency under probing. Expect questions about RAID logs, dependency mapping, resource levelling, stage gates, change control, and what a status report should contain that most contain. Screeners follow up on thin answers.
  • Tooling depth in Excel, PowerPoint, and Word. Not just "proficient" — whether you can spot a broken forecast model or a deck whose narrative contradicts its own data.
  • Rubric discipline. Whether you can apply someone else's grading criteria consistently, including when you personally would have run the program differently.

Logistics

Fully remote and asynchronous. Work is assigned in batches with deadlines rather than fixed shifts; most contributors take 10–25 hours a week and some go heavier during active projects. Pay bands are observed on Mercor listings, not guaranteed — the actual rate depends on the project, your assessed depth, and sometimes a calibration period. Eligibility is limited to the US, UK, Europe, and Canada.