What the work involves
Mercor is building high-fidelity replicas of a marketing org's tool surface: ESP, CRM, product analytics, web analytics, CMS, project management, ads, support inboxes — populated with realistic documents, dashboards, personas and deliberately planted problems. Your job is the Lifecycle slice: contact strategy, milestone campaigns, customer communications, deliverability and list health, and lifecycle portfolio optimization.
Day to day, you write task briefs drawn from work you actually shipped — a suppression rule that quietly killed a winback stream, a milestone campaign double-sending because two triggers overlap, a deliverability slide that is really a list-acquisition problem. Each brief needs the planted issue specified, the tool states and artifacts that make it realistic, a reference answer, and grading criteria that separate a genuinely strong response from a plausible-but-wrong one. Then you review AI agent attempts and score them against your own standard.
What the platform screens for
- Ownership, not oversight. The listing is explicit: you have owned the sending program and its incrementality, not only managed people who did. Screeners probe for holdout design, send-volume decisions, and what you did when a program looked good on open rate and flat on revenue.
- Tool fluency across the stack — an ESP, a CRM such as Salesforce, product analytics such as Amplitude — because environment specs are written in the vocabulary of actual tool states, not abstractions.
- Evaluation judgment. Can you name the error a weaker practitioner would make, and articulate why it is wrong in writing? Much of the value here is the explanation, not the answer.
- Prior experience building AI training environments, RL environments or simulated case studies is strongly preferred, as is Rubric Academy or Rubric Bootcamp certification. Neither is stated as mandatory.
Applicants complete a short multiple-choice knowledge screener specific to lifecycle before human review.
Logistics
Remote and asynchronous, project-based, with hours you choose — most contributors work in blocks rather than fixed shifts. Pay has been observed in the $60–100/hr range depending on depth and the specific workstream; it is not guaranteed. Expect a heavily written workflow: briefs, rubrics and grading rationale are the deliverables.