What the work actually is

You will be handed realistic, messy scenario packets — trial balances, term sheets, purchase and partnership agreements, broker statements, AP aging reports, internal emails — and asked to either produce the professional deliverable yourself or evaluate an AI attempt against what a controller or partner would sign off on. Typical tasks include trader-versus-accounting PnL and mark-to-market reconciliations, vendor payment prioritization under a cash constraint with PO/P-card controls, integrated 3-statement models with ASC 805 deferred-revenue haircuts and PPA amortization, covenant compliance and headroom analysis, and profit-pool waterfalls with capital account roll-forwards. Every packet contains deliberate distractors: stale schedules, immaterial variances, a memo that contradicts the agreement. Separating signal from noise is the skill being measured.

When you are in the evaluation seat, the job is not to say whether the model's answer feels right. It is to identify precisely where the reasoning breaks — a sign error in an intercompany elimination, revenue recognized on a haircut balance that should have been written down, a covenant calculated on reported rather than adjusted EBITDA — and to write the correction so a non-specialist reviewer can verify it. Rationales are part of the deliverable, not an afterthought.

What the screen looks for

  • Hands-on depth, not oversight depth. Screeners push on whether you built the reconciliation or reviewed someone else's. Expect follow-ups on tie-out mechanics, specific journal entries, and what you did when the balance didn't clear.
  • Standards fluency under pressure. US GAAP or IFRS applied to a fact pattern, not recited. ASC 805 and 606 come up often; so does the difference between a judgment call and an error.
  • Written clarity. A short sample of your own explanatory writing carries real weight, because that's the artifact the model learns from.
  • Domain fit. Investing and portfolio backgrounds are explicitly not the target. Product control, FP&A, technical accounting, treasury and P2P, commercial credit, and partnership finance are.

Logistics

Fully remote, contract, asynchronous. Hours are flexible with no fixed shifts; contributors typically commit 10–20 hours a week, though volume moves with project demand and is not guaranteed. Tasks are delivered through Mercor's platform with per-task deadlines rather than live calls. The stated rate is $85/hr as observed on this listing; actual rates can vary by project, assessed skill tier, and geography.