What the work involves

Ethos is staffing this for a foundational AI lab that wants its model to produce documents, spreadsheets, and slide decks a real investor would accept. You'll spend most of your time in one of two modes. In the first, you author gold-standard artifacts from a brief: a Series A deck outline with the story beats in the right order, a data room index structured the way diligence actually proceeds, a quarterly LP report with capital account roll-forwards that tie. In the second, you grade model attempts — marking where the cap table math breaks under an option pool refresh, where a term sheet summary quietly omits the liquidation preference stack, where a deck says "strong traction" without a number behind it — and rewriting the output to the standard you'd send to a partner.

Tasks come with rubrics, but rubrics never cover everything, and the written rationale you attach to a score is often worth more to the lab than the score itself. Expect to explain why a $12M post-money on a $2.5M raise with a 15% pool changes founder dilution, not just that it does.

What the platform screens for

  • Verifiable transaction exposure. Named deal types, stages, fund sizes, and your actual role — not "supported fundraising efforts."
  • Mechanics under follow-up. Pre- vs post-money SAFEs, participating vs non-participating preferred, pro rata, pay-to-play, management fee offsets, DPI/TVPI/RVPI. The voice screen will push one layer past your first answer.
  • Artifact craftsmanship. Whether you can describe what separates a clean model from a fragile one, and what an LP notices on page one of a report.
  • Evaluation judgment. Can you distinguish a stylistic preference from a factual error, and can you rank severity when a document has several problems at once?

Logistics

Fully remote and asynchronous — no standing meetings, no set hours. Volume is flexible between roughly 5 and 20 hours a week, with more available to contributors whose work holds up in review. The $100/hour figure is the rate observed on this listing and is not a guarantee; rates on evaluation platforms vary by task type, review outcomes, and queue. Onboarding includes calibration tasks before paid volume opens up, and the screening stage includes an AI-conducted voice interview.