What the work involves
You'll spend most of your time on two kinds of tasks. The first is production: given a scenario — a federal IT services RFP, a management consulting orals slot, a pricing volume that needs a summary deck — you build the artifact a proposal manager would actually deliver, and annotate the reasoning behind structural choices. The second is evaluation: the model produces an executive summary, a win theme slide, or a compliance matrix, and you score it against criteria the lab supplies, then write the critique that explains precisely where it failed a real evaluator's standard.
The critiques are the valuable part. "This deck is weak" teaches a model nothing; "the win theme is stated as a company capability rather than tied to a customer pain point named in Section L, and the proof point is an award rather than an outcome metric" teaches it a great deal. Tasks arrive in batches through the platform, each with its own rubric, and reviewers send work back when the reasoning is thin.
What the platform screens for
Ethos runs an AI voice screen before onboarding. It probes for bids you personally managed — scope, role, colour team process, outcome — and follows up on specifics when answers stay general. Expect questions on compliance matrix construction, Section L/M or equivalent evaluation criteria mapping, orals coaching, and how you'd distinguish a deck that reads well from one that scores well. Expert-level PowerPoint craft is treated as a hard requirement, not a nice-to-have: slide architecture, action titles, graphic-to-text ratio, and template discipline all come up.
Logistics
- Fully remote, fully asynchronous — no standing meetings
- 5–20 hours per week, self-scheduled, with more available to contributors whose work clears review consistently
- $70/hour observed on this listing; rates on AI evaluation platforms vary by task type and tenure and are not guaranteed
- Work is covered by confidentiality terms; you cannot reuse client proposal content, so bring your judgment rather than your files