What the work involves

You work inside AutoCAD Mechanical the way you already do on the job, while the platform captures your screen and your reasoning. Tasks typically mean executing a realistic drafting brief end to end — building a detail sheet to a drawing standard, applying dimension styles and GD&T callouts, generating a BOM from a parts list, inserting standard fasteners from the content libraries, or fixing a drawing someone else set up badly. Alongside the capture you narrate or annotate: why this layer, why this hole-chart approach, why the balloon order is what it is. Some batches ask you instead to judge an AI agent's attempt at the same task and mark precisely where its command sequence, layer discipline, or standards compliance broke down.

The distinction that matters is between someone who knows the ribbon and someone who knows mechanical drafting. Screening is built around that. Expect questions about AM-specific behaviour — mechanical layers versus plain AutoCAD layers, power dimensioning, the difference between an AM part reference and an ordinary block, how associative detail views behave when geometry updates — rather than generic CAD trivia.

What the platform screens for

  • Live access, verified. You need AutoCAD Mechanical installed and openable today on a Windows machine you control. Mercor does not issue licences, virtual machines, or hardware, and AutoCAD LT or plain AutoCAD does not substitute for the Mechanical toolset.
  • Capture resolution. Screenshots must exceed 2.5 megapixels. A single QHD, 4K, ultrawide, or high-DPI laptop display clears it; so does two 1920×1080 monitors captured as one spanning image. A lone 1080p screen does not.
  • Production history. Drawings released to manufacturing, tooling packages, revision-controlled sheets — work where a mistake cost something.
  • Legible explanation. Tasks are useless if the reasoning is unrecorded or vague.

Logistics

Fully remote and asynchronous. Work is released in batches; contributors commonly take 10–20 hours a week and set their own hours, with no standing meetings. Pay has been observed in the $45–55/hr range and varies by batch complexity and reviewer track — it is not a guaranteed rate. Engagements are contract-based and volume fluctuates, so treat this as a supplement to practice rather than a replacement for it.