What the work actually is

You build scenarios that look like the messy middle of product work: conflicting user research, an engineering estimate that blew up, a exec who wants the feature next quarter, a metric that moved for the wrong reason. Each scenario runs from the ambiguous problem statement through to a shipped decision, and you write the reference answer that makes the prioritization logic explicit — not just what you'd cut, but the cost of cutting it and the signal you'd watch afterward. The second half of the job is grading: a model produces a spec, a roadmap, or a strategy memo, and you judge whether the reasoning holds. The failure mode you're hunting is the plan that reads beautifully, cites the right frameworks, and quietly never chooses.

What the screen looks for

AfterQuery's screening is AI-led and leans hard on specifics. Expect to be asked which product you owned, what you shipped, what the numbers did, and what you got wrong — then to be asked a follow-up that assumes you were actually there. Vague ownership language ("I drove alignment across stakeholders") collapses under a single probe about who disagreed with you and how it resolved. The evaluation side is tested separately: you'll likely be given a plausible-sounding product artifact and asked what's wrong with it, which measures whether you can articulate why a tradeoff was dodged rather than just noting that you'd have done it differently.

Logistics

  • Fully remote and async — no standups, no meetings on the calendar
  • Minimum commitment of 10 hours per week; you set when they happen and can scale week to week
  • Weekly pay via Stripe; the observed band is $60–110/hr, set from your experience and not guaranteed at any particular point in the range
  • Work is judged on judgment quality, not task volume, so throughput padding doesn't help you

Depth in B2B SaaS, consumer, or platform product is welcome, as is a technical background or long stretches working shoulder to shoulder with engineers. Prior experience with AI or data products is a plus but not a gate — most people who do well here have simply owned something real and can defend the decisions they made on it.