What the work actually involves
You will spend most of your time reading model output that looks professional and deciding whether it holds up. A typical task might be a DCF with a terminal growth rate that quietly exceeds the discount rate, a comps set that includes a company with a different capital structure, or a buy recommendation whose supporting math contradicts the thesis paragraph above it. You flag the error, explain why it matters to an investment decision, and write feedback specific enough that a model trainer with no finance background can see the failure. A second common format is head-to-head comparison: two responses to the same prompt, and you rank them on analytical rigor, valuation soundness, and the quality of the recommendation — then justify the ranking in writing.
What the platform screens for
Mercor's screen is AI-led and leans hard on verifiable specifics. Expect to be asked what you valued, using which method, at which firm, and what the output was used for. Vague claims about "financial modeling experience" invite follow-ups that quickly separate people who built models from people who read them. The screen also probes evaluation judgment — whether you can distinguish an error that changes the investment conclusion from a stylistic preference, and whether you can articulate a standard rather than a reaction. Written English matters more than usual here, because the written rationale is the deliverable.
Logistics
- Fully remote and asynchronous; you pick up tasks from a queue rather than attending meetings.
- 20+ hours per week is a stated requirement, not a suggestion — throughput is tracked.
- Duration is roughly 4–6 weeks with an immediate start.
- Payment structure has two stages: the first task is paid on approval, and completing it within the required timeline qualifies you for hourly pay for the remainder. Rates in the $70–110/hr band are observed on comparable Mercor finance projects, not guaranteed for any individual.