Mercor is building simulated marketing-org environments — CRM, marketing automation, project tracker, support desk, analytics — and needs marketing ops practitioners to author the tasks inside them and grade AI agent attempts. You design briefs from work you've actually done, plant the failure mode, specify the tool states and documents, write the reference answer, and judge agent transcripts against your own standard.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work actually involves

Each environment reproduces a marketing org's tool surface — email, storage, Salesforce or equivalent CRM, marketing automation, Asana or Linear, Zendesk or Intercom, web and product analytics, ads, CMS — populated with realistic records, dashboards, personas and deliberately planted problems. Your job is to author the marketing operations tasks that live in it: a lead routing rule that silently drops a segment, a campaign UTM taxonomy that breaks attribution reporting, an agency handoff where the documented process and the actual workflow state disagree, a reconciliation between CRM opportunity counts and the automation platform's send logs.

A task is not just a prompt. You specify what the task is really testing, the planted issue, and the decision a strong practitioner should reach — then the artifacts required to make that discoverable: which objects exist in the CRM, what the dashboard shows, which Slack thread or ticket contains the contradicting detail, what a stakeholder persona will and won't volunteer. You write or review the reference answer and the criteria that separate a strong response from one that is plausible but wrong. Then you read AI agent attempts and grade them against that standard, in writing.

What the platform screens for

Applicants complete a short multiple-choice knowledge screener specific to marketing operations before human review. Beyond that, the screen is looking for hands-on build history — you configured the routing logic, the lifecycle stages, the sync mappings and the reporting yourself, rather than managing someone who did. Expect follow-ups that go a layer deeper than your first answer: which tool, which object, what the field-level behaviour was, what broke and how you found it. Prior experience building AI training environments, RL environments or simulated case studies is strongly preferred, as is Rubric Academy or Rubric Bootcamp certification. Written reasoning carries real weight here — much of the deliverable is the explanation of why a decision is right.

Logistics

Fully remote and largely asynchronous; work is claimed and delivered in units rather than shifts
Observed band of $60–100/hr, typically scaled to depth of experience and task type; not guaranteed
Part-time volume is normal — many contributors run this alongside a full-time role
Turnaround expectations are per-task, so realistic weekly capacity matters more than time zone

Marketing Expert — Marketing Operations (AI Agent Environments)

What the work actually involves

What the platform screens for

Logistics

Interview prep — 10 questions