Finance Specialist

$65–90/hrrate as listed

Write and grade finance tasks that train frontier language models, working as a W-2 employee of Cincinnatus LLC placed inside a leading AI lab's GenAI team. The role requires 8+ years at a top-tier finance institution and prior hands-on experience scoring model outputs against rubrics.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

You will sit alongside research and engineering teams as the finance voice in the room. Day to day, that splits into two halves: authoring and evaluating. On the authoring side, you design finance problems hard enough to expose where a model's reasoning breaks — valuation under messy assumptions, credit analysis, capital structure decisions, deal mechanics, portfolio construction — and write the reference solution with the reasoning fully shown, not just the answer. On the evaluation side, you score model responses against structured rubrics and write feedback that explains why a response is wrong, which is usually harder than spotting that it is wrong.

A meaningful portion of the job is rubric work itself: proposing scoring criteria for finance tasks, arguing for what should count as a material error versus a stylistic one, and reconciling disagreements with other SMEs so labels stay consistent across the dataset. Expect calibration discussions and periodic revisions to guidelines as the lab's priorities shift.

What the platform screens for

Mercor's screen is AI-led and leans on verifiable specifics. Expect probing on where you worked, what you actually owned, and how your title progressed — Analyst to Associate to VP or Director is called out explicitly in the listing. The prior LLM-evaluation requirement is stated as mandatory, so be ready to name the platform or program, describe the rubric dimensions you scored on, and give a concrete example of a judgment call you made. Generic familiarity with ChatGPT will not clear this bar. Domain follow-ups tend to go one or two layers past the first answer, so choose examples you can defend in detail.

Logistics

W-2 employment through Cincinnatus LLC, which acts as employer of record and handles payroll, benefits, and compliance
Placement is with the client AI lab's extended workforce
Stated commitment is at least 35 hours per week, weekdays — this is close to full-time and not a nights-and-weekends side engagement
Remote, with collaboration and calibration work that assumes reasonable overlap with weekday business hours
Observed pay band for comparable finance SME roles on this platform runs $65–90/hr; actual offers vary with seniority and are not guaranteed

What they look for

8+ years in finance at a recognized top-tier firm (IB, asset management, corporate finance, or Big Four advisory)
Hands-on experience evaluating LLM outputs against rubrics or structured scoring criteria
Demonstrable title progression (Analyst → Associate → VP/Director)
Ability to write clear, well-reasoned written solutions and critiques
Availability for 35+ hours/week on weekdays
Comfort collaborating with other SMEs on rubric design and label consistency

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Do you have 8 or more years of professional experience in finance roles (e.g., investment banking, asset management, corporate finance, financial advisory)?
Years of professional finance experience · required
Have you worked at a recognized, top-tier financial institution such as Goldman Sachs, JPMorgan, Morgan Stanley, BlackRock, Fidelity, Deloitte, PwC, EY, KPMG, or equivalent?
Top-tier organization experience · required
Do you have hands-on, prior experience evaluating LLM or AI model outputs against rubrics or structured scoring criteria?
LLM evaluation experience · required
Can you reliably commit to at least 35 hours per week during weekdays?
Weekday availability · required

0 of 4 answered

Included with this listing

Interview prep — 12 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether the candidate has genuinely evaluated model outputs before, not just talked about finance—and whether they can articulate the judgment calls that come up when scoring reasoning, not just checking boxes on a checklist.
A strong answer: A strong answer names a concrete project (not hypothetical), describes the financial domain tested (e.g., valuation, risk assessment, M&A analysis), explains the rubric structure (correctness, reasoning quality, completeness), and shows how they resolved ambiguity—e.g., when a model's recommendation was sound but the math had a minor error, or when multiple defensible approaches existed.
Where people lose it: Claiming evaluation experience but describing only passive review of outputs, or offering only vague reference to 'assessing AI quality' without explaining the rubric, the financial specifics, or how they actually scored anything.