What the work actually involves

You are assigned a lifestyle scenario you have genuinely lived — maintaining a training log across an injury, running a 120-guest wedding, planning a multi-city road trip with a dog in the car — and you turn that practice into a machine-usable recipe. That means writing the domain rules and gotchas in your own words (why a race taper week breaks a naive weekly-volume average; why caterer headcount deadlines fall before RSVP deadlines and what people do about it), drafting the interview questions the assistant should ask a user before it builds anything, and constructing a clean multi-tab Google Sheet that acts as the golden reference for the data model and formulas. You then test the visual app the assistant generates from your recipe and iterate until it reflects how the task really works, signing off through daily check-ins and QC review.

What the platform screens for

Two things carry most of the weight. First, lived specificity: screeners probe for the edge cases that only show up after you've done something repeatedly, and generic advice available in any blog post reads as a fail. Second, authentic human authorship — the client evaluates every deliverable for human voice, and AI may only polish syntax, never supply domain insight. Beyond that, expect questions on Google Sheets structure (multi-tab layout, everyday formulas, data validation) and on assistive logic: whether you can predict what a vague user request leaves unsaid and design a short interview that resolves it without interrogating them.

Logistics

  • Fully remote, contractor engagement, up to roughly three weeks per sprint
  • Stated commitment of 40 hours per week with 8 hours of overlap with PST; sprint effort described as 35–50 hours with daily check-ins
  • Weekend on-call availability required; part-time engagement is acceptable
  • Your own desktop or laptop and a stable high-speed connection
  • Pay band is not disclosed in this listing; ask for the rate and sprint scope on the Job Interest Form

Shortlisted candidates receive a Job Interest Form, then finalists go through a delivery interview. A bachelor's degree or equivalent practical experience in any field is listed; prior AI evaluation, annotation, or QA work is preferred but not required.