What the work involves

You will either conduct structured elicitation sessions or assess recordings and transcripts of sessions conducted by others, including model-generated interviewers. Day to day that means building rapport without pressure, adjusting your line of questioning as the subject's tone, pacing and hesitation shift, and then stepping back to write up what you concluded and why. Every read carries a confidence assessment and a traceable rationale: not "the subject seemed evasive" but which answer contradicted which, where the latency changed, and what alternative explanation you ruled out.

A significant share of the work is rubric scoring. You will annotate sessions against a detailed rubric — internal consistency, reliability, evasiveness indicators, emotional and motivational drivers — and your scores need to hold up against other collectors working the same rubric. Consistency across twenty cases matters more than brilliance on one.

What the platform screens for

micro1 runs an AI-led interview before any human sees your file. It probes for specifics: which methodology you were trained in and by whom, how many interviews you have actually run, whether you can name a concrete case where you revised an initial read and what moved you. Claims of intuition without articulable basis score poorly — the whole point of the project is externalizing the reasoning. Expect follow-ups that push on a single answer two or three levels deep, and expect at least one scenario where you must justify a confidence rating you would rather hedge on.

No AI or machine-learning background is required or expected. Familiarity with after-action reporting or annotation work is a plus, not a gate.

Logistics

  • Contractor engagement, roughly 15+ hours a week, with hours you largely set yourself
  • Remote; US-based candidates preferred, partly for session scheduling overlap
  • Observed band of $25–65/hr, varying with depth of experience and whether you are conducting or assessing sessions — stated as observed, not guaranteed
  • Live sessions require scheduled availability; scoring and annotation work is asynchronous
  • Reliable internet, a quiet space and a decent headset are practical necessities for any recorded or live interaction