What the work involves

You are not inspecting boards for a manufacturer — you are checking whether an AI system inspects them correctly. Typical tasks include reviewing a Gerber or CAM package and rating a model's manufacturability assessment; looking at an AOI flag and deciding whether the model correctly separated a true solder defect from a false call; reading a model's IPC-A-610 Class 2 versus Class 3 acceptance judgment and marking exactly which acceptance criterion it misread; or reviewing its interpretation of a C-SAM delamination image or X-ray void measurement against MIL-STD-883 method language. Some work is generative — writing the reference answer, the defect taxonomy, or the counterfeit risk-indicator rationale under AS6081 that the model will be trained against.

The written rationale is the deliverable. A rating with no explanation has almost no training value, so reviewers are expected to cite the standard, the class, the paragraph or method where practical, and the physical mechanism — voiding from outgassing, tombstoning from unequal pad thermal mass, insulation crush from an over-set crimp die. Assignments usually arrive in batches with a rubric and a calibration set; disagreement with another reviewer is adjudicated rather than averaged, so being able to defend a call matters.

What the screen looks for

micro1's process is AI-led: a conversational interview that asks for your background, then probes a specific domain area with follow-ups until it finds the edge of your knowledge. It rewards concrete detail — tools you ran, part families, failure modes you actually diagnosed, which IPC class your product lines were built to — and penalises fluent generalities. Certification (CIS/CIT) is checked, not assumed. Expect at least one scenario where the model's answer is plausible but wrong, and your job is to say why.

Logistics

  • Contractor engagement, fully remote, US-based preferred.
  • Asynchronous task queues; most reviewers work 10–25 hours a week, some full-time during ramp periods.
  • Observed rates for this listing are $70–120/hr, set by seniority and task type — stated as observed, not guaranteed.
  • No prior AI or machine-learning experience expected; domain depth and clear writing are the currency.