What the work actually involves

You write procurement problems that a capable AI agent should be able to solve, and then you write the answer key. A typical task might be: five supplier quotations for a mid-size IT infrastructure refresh, arriving in a PDF, two spreadsheets with different line-item granularity, an email body, and a scanned specification sheet — none of them comparable without normalisation. You build the artifacts, define the defensible recommendation, and document why the second-cheapest bid wins once total cost of ownership, SLA credits, and non-conforming exclusions are factored in. Then you build a rubric granular enough that an AI grading agent can distinguish a model that reasoned correctly from one that landed on the right vendor by accident.

The rubric writing is usually harder than the task writing. You need to decide what partial credit looks like when a model spots the freight-terms discrepancy but misses the currency exposure, and how to score a recommendation that reaches a different conclusion via sound logic. Expect review cycles with micro1's project team, and expect your rubrics to be pushed back on when they encode a preference rather than a defensible procurement standard.

What the platform screens for

  • Verifiable procurement tenure — roughly four or more years in corporate or public-sector buying, with hands-on ownership of supplier analysis, bid matrices, and negotiation, not adjacent finance or logistics work.
  • Depth in a category — IT, industrial, or professional services. Screeners follow up on specifics: pricing models, common vendor tactics, where quotes are deliberately made non-comparable.
  • Regulatory and methodological grounding — public procurement thresholds, evaluation frameworks, scoring weightings, conflict-of-interest handling.
  • Written English at documentation standard, since much of the deliverable is prose justification.

No AI or machine-learning experience is required or expected.

Logistics

Remote contractor engagement, largely asynchronous, with periodic calls for task calibration and rubric review. Observed pay for this listing sits in the $30–65/hr range — stated as observed, not guaranteed, and typically tied to category depth and demonstrated rubric quality. Hours are usually flexible and project-dependent; some contributors work part-time alongside a primary role, others take heavier task volume during ramp periods.