What the work involves

You will spend most of your time on two kinds of tasks. The first is authoring: given a realistic prompt — a Series C board update, an all-hands narrative after a reorg, a QBR package for a regional sales org — you produce the deck or the written communication the way you would have produced it for your CEO, including the structure decisions, the page-level argument, and the speaker-note logic that makes a slide defensible in the room. The second is evaluation: you are handed model-generated attempts at the same brief and asked to judge them, rank them against each other, and write the rationale explaining why one framing survives a board's questions and another collapses. The rationale is the product. A rating without a legible reason is nearly worthless to the lab.

Expect the work to surface the parts of the job that are rarely written down: when an appendix protects the narrative, how much bad news goes on the page versus in the pre-read, why a metric is indexed rather than absolute, what a compensation committee will read first. Models are usually competent at producing something that looks like a board deck and weak at the judgment underneath it. Your job is to make that judgment explicit.

What the platform screens for

Ethos runs an AI voice screen before any project assignment. It probes for real tenure supporting C-level principals at companies with actual governance — a board, a real planning cycle, external reporting obligations — and it follows up. Expect questions about specific decks you owned end to end, how you handled a slide your CEO disagreed with, and what you cut from a board pre-read and why. Vague seniority claims tend to unravel two follow-ups in. The screen also tests evaluation temperament: whether you can articulate a standard rather than just assert taste, and whether you can be decisive about a genuinely close call between two adequate drafts.

Logistics

  • Fully remote and asynchronous; no standing meetings, no client-facing calls.
  • Flexible volume, typically 5–20 hours per week, with more available to contributors who deliver consistently.
  • Work is delivered through the platform's task interface; you will need PowerPoint (or equivalent) and a workflow that lets you produce slides, not just critique them.
  • Pay is observed at $80/hour for this role; rates are set per project and are not guaranteed.
  • Confidentiality matters — you will be asked to generalize from real experience without reproducing any employer's actual materials.