What the work involves

You build self-contained statistical scenarios and the reference reasoning that goes with them. A typical item starts from a plausible applied setting — a staggered rollout of a pricing change, a clinical endpoint with informative dropout, an A/B test where randomization is compromised by shared devices — and asks a model to choose an estimand, justify an identification strategy, run the analysis, and interpret the result. Your reference answer has to make the reasoning explicit: why difference-in-differences rather than a naive pre/post comparison, what the parallel trends assumption buys you, when a clustered standard error changes the conclusion.

The second half of the job is grading. You read model output that is fluent, correctly formatted, cites the right method names, and is nonetheless wrong — a p-value interpreted as the probability the null is true, a peeking-corrupted sequential test reported as significant, an instrument justified by relevance while exclusion is quietly assumed. Flagging analyses that read rigorous but draw invalid conclusions is the explicit deliverable, and it is the skill the labs are actually paying for.

What the platform screens for

AfterQuery's screen is AI-led and follow-up heavy. Expect it to take a method you name and push one layer deeper: state your assumptions, tell it what breaks them, tell it how you would detect the break in data you actually had. Generic familiarity with a technique surfaces quickly. Bring two or three concrete pieces of applied work you can narrate end to end — the design, the decision you argued for, the analysis you ran in R or Python, and what the result changed. Advanced degree plus two years of applied experience is the stated floor; PhDs and people who have operated experimentation platforms at scale tend to land higher in the $80–150/hr band, though rates are as observed and not guaranteed.

Logistics

  • Fully remote, fully asynchronous — no standing meetings, no fixed timezone
  • Minimum commitment of 10 hours per week, with the schedule set by you
  • Paid weekly via Stripe
  • Ongoing project work rather than a fixed-term engagement