What the work involves

Ethos is staffing a foundational AI lab's effort to make its model competent at professional document, spreadsheet, and deck work in a revenue operations context. In practice you will do three kinds of tasks, often in the same session. You write reference artifacts from a realistic prompt — a CRM field-mapping schema between a marketing automation platform and Salesforce, a lead-routing logic document with fallback rules, a territory carve model with account-count and pipeline-coverage balancing, a quarterly pipeline hygiene audit, a funnel conversion dashboard spec, or a deduplication ruleset with match keys and merge precedence. You grade model attempts at the same tasks against a rubric. And you rewrite weak model output into something you would actually forward to a VP of Sales, with notes explaining each change.

The judgment being captured is the unglamorous kind. A routing document that looks tidy but has no rule for an inbound lead with a blank country field is wrong in a way that matters. A funnel dashboard that computes stage conversion on a snapshot basis when the question was cohort-based is wrong in a way that misleads leadership for a quarter. Your value to the lab is catching that and saying precisely why.

What the platform screens for

  • Verifiable hands-on ownership. Five-plus years in revenue, GTM, or CRM data operations, where you personally owned schemas, routing logic, or dedup rules rather than briefing someone who did.
  • Depth under follow-up. The AI voice screen asks a question, then asks a harder one built on your answer. Naming tools is not the same as explaining a merge-precedence decision and what broke when you got it wrong.
  • Artifact craft. Spreadsheet modeling, slide construction, and written documentation that a sales leader can act on without a follow-up meeting.
  • Evaluation instinct. Whether you can articulate why one output is better than another in terms a rubric can carry — not just flag that something feels off.

Logistics

Fully remote and fully asynchronous. Commitment is flexible between 5 and 20 hours per week, with more available if you want it; there are no standing meetings and no fixed hours. Task volume varies by week, so treat the band as a range rather than a guarantee. The interview process includes an AI voice screen. Pay is $100/hour as observed on this listing, not a guaranteed rate.