What the work actually involves

You write tax problems the way they show up in practice, then judge whether a model handles them like a competent preparer or reviewer would. A task might mean drafting a fact pattern for a multi-state 1065 with special allocations and a §754 election, preparing the correct treatment yourself, and then scoring two model responses on technical accuracy, authority citation, and whether the position taken is defensible on exam. Other days you'll be reviewing an ASC 740 provision walkthrough — rate reconciliation, valuation allowance judgment, uncertain tax positions — or a response to an IRS notice or state IDR, and writing feedback specific enough that a researcher who is not a tax person understands exactly where the reasoning broke.

The hardest part is usually not spotting the error. It's articulating why the wrong answer was tempting — the Code section that looks like it applies but doesn't, the election deadline that changes everything, the difference between a position that is right and one that is merely arguable.

What the platform screens for

  • A real, provable credential. Active CPA or EA, with license state and status generally verified.
  • A named sub-specialty. Generalist answers screen poorly. Corporate, individual/HNW, partnership and pass-through, SALT, or international/cross-border — pick one and go deep.
  • Depth under follow-up. The AI interviewer will push past your first answer on things like basis tracking, apportionment methodology, GILTI/FDII mechanics, or §382 limitations. Vague answers surface fast.
  • Written clarity. Feedback quality is the product here. Terse, cited, unambiguous writing beats volume.
  • Judgment on gray areas. Much of tax has no single right answer; screeners look for whether you can distinguish substantial authority from wishful thinking.

Logistics

Fully remote, asynchronous, contractor engagement. Most contributors work 10–20 hours per week around a day job, though some projects offer more volume during ramp periods. Work is claimed from a queue rather than scheduled, so evenings and weekends are viable. Pay in the $80–120/hr band has been observed on comparable Mercor tax projects and varies with sub-specialty scarcity, credential, and calibration performance — it is not guaranteed. Selected applicants typically complete a short paid or unpaid sample task before onboarding.