What the work involves

This is presentation craft as training data. Ethos routes you tasks drawn from the real SE/SC workflow — a technical solution deck for a platform evaluation, a demo storyboard, an executive-facing architecture diagram, a POC results readout, RFP response slides, value engineering exhibits. Some tasks ask you to build the gold-standard artifact yourself. Others hand you a model-generated deck and ask you to rank it against alternatives and write out what's wrong: a five-box architecture diagram that would be right for engineers and fatal for a CFO, a POC summary that buries the one number the buyer cares about, a value slide with unsourced ROI math.

The written critique usually matters more than the score. A reviewer who writes "slide 4 is weak" produces nothing usable; a reviewer who writes "slide 4 leads with integration architecture when the stated audience is a procurement committee — the latency benchmark on slide 9 should be the opening claim, and the architecture belongs in an appendix" is producing the signal the lab pays for.

What the platform screens for

  • Real SE/SC tenure. Four-plus years carrying technical pre-sales at a recognized technology company. Expect follow-ups on deal stage, audience, and what you actually built versus what marketing handed you.
  • Audience calibration. Whether you can articulate the concrete differences between a deck for a staff engineer, a VP of engineering, and an economic buyer — and change a slide accordingly.
  • Slide mechanics. PowerPoint at an expert level: masters, alignment discipline, build sequencing, chart selection, what belongs in the appendix.
  • Written judgment. The AI voice screen probes how you'd justify a ranking out loud, because that reasoning becomes your written rubric later.

Logistics

Fully remote and asynchronous. Five to twenty hours per week, more if you want volume and your quality scores hold. No standing meetings; work is claimed from a queue and turned around on your own schedule. The observed rate on this listing is $80/hour — banded by the platform, not guaranteed, and in practice tied to sustained review quality rather than raw throughput. Screening starts with an AI voice interview.