What the work involves
You'll read AI-generated military reporting products and judge whether they would survive a staff review. That means checking whether a Weekly Activity Report actually rolls up inputs the way a headquarters section would submit them, whether an After-Action Report separates observation from discussion from recommendation, whether the terminology is current doctrine or a plausible-sounding hallucination, and whether an executive summary tells a General Officer what they need in the space they'd actually read. Alongside evaluation, you'll write realistic reporting examples and scenarios — sanitized, unclassified — that give the model something correct to learn from.
Most tasks are asynchronous: a queue of outputs, a rubric, and a written-feedback field. The feedback is the deliverable. "This is wrong" earns nothing; "the AAR conflates a sustain item with a recommendation, and the recommendation has no assigned OPR or suspense, which is what makes AARs actionable at brigade level" is the standard.
What the platform screens for
micro1's screen is AI-led and probes for firsthand staff experience rather than general familiarity with the military. Expect questions about which specific recurring reports you personally built, who consumed them, what the battle rhythm looked like, and how you handled a section that missed its input. Vague answers get follow-ups until they either resolve into specifics or don't. No AI or annotation background is required, and saying so is not a disadvantage.
One thing to be clear about up front: everything you produce must be unclassified and releasable. Candidates who volunteer operational detail they shouldn't are a liability, not an asset.
Logistics
- Fully remote, contractor engagement, no fixed shift
- Async task queues; typical commitments run 10–20 hrs/week, with some projects asking more during ramp
- Observed pay $40–80/hr, varying by seniority and task type — not a guaranteed rate
- Work volume is project-driven and can pause between customer phases