What the work actually involves

You are handed a workflow brief — set up a class section, distribute a differentiated assignment, return graded work with rubric-based feedback, reorganize a stream by topic — and you carry it out inside a live Google Classroom instance while recording your screen. The recording is only half the deliverable. The other half is written: a step-by-step account of what you clicked and why, the names of the features used, a realistic prompt a teacher might have issued to produce this task, and a detailed rubric describing what a correct completion looks like. Tasks are self-contained and reviewed against project guidelines, so sloppy narration or a workflow that skips a step invisible on screen comes back for rework.

What the platform screens for

  • Real classroom mileage, not familiarity. Screeners probe for specifics: how you handle a resubmitted assignment, what the difference between returning and draft-grading is, how rubrics attach and score.
  • Ability to narrate while doing. Many strong teachers freeze when recording. micro1 checks that you can talk through intent, not just execute.
  • Documentation discipline. Naming features correctly and writing rubrics that another annotator could apply without asking you a question.
  • Independent access. You need an account you can create classes in — a personal Gmail-based Classroom works for most tasks, but you should be able to state what you have.
  • Prior human-data work (annotation, RLHF, model response evaluation) is treated as a plus, not a gate.

Logistics and pay

Fully remote, contractor, asynchronous. Work arrives in batches; there are no fixed shifts, but batches often carry turnaround windows measured in days, and availability questions during screening are real — inconsistent throughput is the most common reason contributors stop receiving tasks. Observed pay on this listing runs $30–100/hr, with the range reflecting task complexity and how much documentation each workflow demands; rates are set per project and are not guaranteed. Expect some unpaid or low-paid time in onboarding and calibration before volume stabilizes.