What the work involves
You will spend most of your time doing two things: producing reference-quality slides yourself, and judging slides the model produced. A task might hand you a raw data table and a client brand system and ask for a finished exhibit; another might show you a model-generated waterfall chart and ask you to mark every place it violates alignment, labelling, or hierarchy conventions — then rewrite the instruction that would have produced a correct build the first time. Common task families include master template and layout design, exhibit and chart formatting, visual hierarchy on dense pages, brand-compliant deck production, complex graphic builds (process flows, org charts, timelines, Gantts), and fast turnaround fixes on broken decks.
Written rationale matters as much as the artefact. A deck you rate as weak is only useful to the lab if your reasoning is specific enough that a model can learn from it: not "looks unprofessional" but "three type sizes on one page with no hierarchy rule, footnote markers unattached to the figures they qualify, and a legend duplicating series labels already on the bars."
What the platform screens for
Ethos runs an AI voice screen before any task access. It probes where you built slides, at what volume, under what turnaround expectations, and whether you worked inside a real brand system or improvised per deck. Expect follow-ups that get concrete fast — how you'd fix a misaligned stacked bar, how you handle a partner's markup that conflicts with the template, what you actually do in master slide layouts versus on individual pages. Vague seniority claims collapse under those follow-ups; specific production stories hold up.
Logistics
- Fully remote, asynchronous, no fixed hours or standing calls
- Flexible 5–20 hours per week, with more available to consistently strong contributors
- $70/hour is the rate observed on this listing, not a guarantee; volume varies with lab demand
- Work is self-scheduled, but tasks often carry per-batch deadlines