What the work involves

You receive an artifact an AI produced in response to a finance brief: a three-statement model in Excel, a valuation summary deck, a credit memo, an investor update, a comps table. Your job is to judge whether it would survive contact with a real workflow — would an associate send it up, would a client accept it, would a reviewer catch something embarrassing. You grade against a rubric and then write the part that matters: specific, located, structured feedback explaining what is wrong and what correct would have looked like.

Errors come in three flavours and all three count. Factual and analytical: a discount rate that doesn't match the stated capital structure, a balance sheet that doesn't balance, a multiple applied to the wrong metric, a hardcode buried in a formula column. Presentation: a chart axis that distorts the trend, inconsistent units, a slide title that asserts something the chart doesn't show. Usability: a model with no assumptions tab, a deck that buries the recommendation on slide nine. Strong evaluators separate "wrong" from "defensible but unconventional" and say which one they're flagging.

What the screen looks for

Mercor's interview is AI-led and pushes on follow-ups. It wants five-plus years at a named firm in the US, UK, Canada, Australia or New Zealand, and it will ask what you actually built — deal types, model structures, who reviewed your work. Generic seniority claims collapse fast under "walk me through the specific tab." It also tests whether you can critique without rewriting: evaluators who respond to a flawed artifact by describing their own preferred version, rather than diagnosing the one in front of them, score poorly. Excel and PowerPoint fluency is assumed and probed concretely.

Logistics

  • Remote, hourly, asynchronous — tasks are claimed from a queue rather than scheduled
  • Observed band for this listing is $100–150/hr, set by experience and task type; not guaranteed
  • Most contributors work 10–20 hours weekly, though volume fluctuates with project demand
  • Expect an onboarding calibration round before paid work opens up