What the work involves
You'll be working with a foundational AI lab, through Ethos, on the specific problem of making a language model competent at professional presentation tasks. In practice that means three kinds of task, rotating: producing reference decks from a realistic brief (a Q3 all-hands, a reorg announcement, a crisis holding statement with speaker notes), evaluating model-generated decks against a rubric and writing the reasoning behind your score, and rewriting weak model output into something you'd actually put in front of a CEO.
The decks in scope mirror what a comms manager owns in-house: town hall and all-hands narratives, executive talking points paired with slides, crisis communication materials, internal campaign launches, and leadership offsite packages. Your judgment about what belongs on a slide versus in speaker notes, how much a CEO will actually say out loud, and when a message needs legal review is the thing being captured.
What the platform screens for
- Four or more years in corporate comms at a company of real size — the screen probes for audience scale, approval chains, and who signed off on your work.
- Ownership, not support. Building slides someone else scoped is different from owning the narrative. Expect follow-ups on whether you wrote the message or formatted it.
- Slide craft you can describe. Master slides, template governance, chart choices, density decisions. Vague claims of "expert PowerPoint" get pushed on.
- Written articulation of judgment. Most of the pay is for the explanation attached to your rating, not the rating itself.
Logistics
Fully remote and asynchronous — tasks queue and you pull them when you have time. Commitment is flexible at 5–20 hours per week, with more available if you want it. The observed rate is $70/hour; bands on platform work shift with task type and are not guaranteed. Screening includes an AI voice interview, typically 20–30 minutes, that follows up on whatever you claim rather than reading from a fixed list.