What the work involves
You evaluate how well a large language model handles the accounting problems a working practitioner sees every week. That means reading model outputs and judging whether they are technically correct, appropriately cited, and free of the confident-sounding errors that would fail a review. Typical tasks include:
- Scoring model answers on revenue recognition, lease accounting, deferred tax, consolidation, and impairment questions under IFRS or US GAAP
- Writing reference solutions and worked journal entries the model is trained against
- Building test cases from realistic scenarios — a messy reconciliation, an ambiguous accrual, a related-party disclosure question
- Comparing two model responses and explaining, in writing, which is better and precisely why
- Flagging hallucinated standard references (a fabricated ASC or IAS paragraph is a common failure mode)
What the platform screens for
micro1 runs an AI-led interview before any human review. It probes your credential (jurisdiction, license status, firm or industry background), then pushes into technical depth with follow-up questions — expect to be asked why a treatment applies, not just which one. Screens tend to favor candidates who can reason out loud, name the standard they are relying on, and say plainly when a fact pattern is underspecified. Vague or padded answers read poorly. Written English matters, since most of the deliverable is written rationale.
Logistics
Fully remote and largely asynchronous, with work claimed from a queue rather than assigned on a fixed schedule. Most contributors treat it as part-time — commonly 10–20 hours a week — and volume fluctuates with project cycles, so it is better understood as supplemental income than a replacement for practice work. The $30–35/hr band reflects rates observed on this platform for credentialed accounting reviewers; actual offers vary by project, jurisdiction, and specialization and are not guaranteed.