MarieRosa
For executives

Get paid to judge how AI handles executive decisions.

AI labs pay senior operators to grade and improve how their models reason about strategy, hiring, capital and crisis. Marie&Rosa brings that work to women and non-binary executives. The work is remote and the hours are yours.

Who qualifies

We accept current and former executives at VP level and above. You qualify if at least one of these describes you:

  • You hold or held a VP, SVP, EVP or C-suite title with P&L, budget or functional leadership responsibility.
  • You serve or served on a corporate board.
  • You founded a company that reached [revenue threshold].
  • You are or were a partner at a professional services, investment or law firm.

You also need at least 10 years of professional experience. In the first cohort we accept experts based in the United States.

This is a poor fit if you are earlier in your career, if you want a full-time role, or if your experience is mainly individual contributor work. We would rather tell you now.

What the work is

You read what a model says about a business decision and judge it the way you would judge a memo from a direct report. The tasks come in five forms:

EvaluateScore model answers to strategy and leadership scenarios, and explain your score.
RankOrder two or more answers to the same question from best to worst.
Write rubricsDefine what a strong answer looks like for a given decision.
Design scenariosDraw decision scenarios from the kind of situations you have handled.
Red-teamPress a model on the judgment calls where it fails.
Scenario evaluation

A model is asked whether a $40M-revenue software company should cut its sales team by a fifth to extend runway or raise a bridge round. You score its answer on whether it asked the right questions, weighed the people consequences and reached a defensible recommendation. Then you write three sentences on what a seasoned CFO would have done differently.

Rubric writing

A lab needs to grade answers about handling a board that has lost confidence in a CEO. You write the criteria that separate a good answer from a dangerous one, with examples of each.

These examples are illustrations written by Marie&Rosa. They are not client projects.

Time

Five to 15 hours a week, remote.

Most projects are flexible. You choose which projects to accept, and declining one does not affect future offers.

Confidentiality

Your judgment, never your employer's data.

You will never be asked to share a current or former employer's confidential information, and you must not use it. Each task asks you to confirm this. The work draws on your judgment and experience, and on nothing your employer owns.

What we keep: the details we need to verify you and pay you. What we never ask for before acceptance: tax forms, identity documents or bank details. The full terms are in our privacy policy and expert terms.

Pay

Paid by the hour or by the deliverable.

Proposed bands, pending approval
Proposed hourly pay by experience and type of work
Your experienceEvaluating and rankingRubrics, scenarios and red-teaming
VP, SVP and functional heads$125 to $175 an hour$150 to $200 an hour
C-suite, P&L owners, founders above [revenue threshold], board directors$175 to $250 an hour$200 to $300 an hour
Sitting or former CEOs of large companies, public company directors$250 to $400 an hourPriced by engagement

Some work pays per deliverable instead. A decision scenario with its rubric pays $400 to $1,000 depending on scope.

You are paid [payment schedule, for example within X days of approved work] as an independent contractor. We collect tax forms only after you are accepted.

For context, Mercor reports an average contracted rate of $90 an hour across all its experts.1

How to apply

  1. Email and LinkedIn URLUnder a minute.
  2. Confirm your detailsYou can sign in with LinkedIn so your name and email fill in, or enter everything by hand. If you paste your profile URL, we show you every field we pulled and you confirm or edit each one before anything is saved.
  3. Identity and title checkWe confirm you held the role you list.
  4. A 20-minute conversationWith Tran or a designated reviewer, plus one short written exercise: one strategy scenario and one model answer to evaluate. It is a sample of the real work.
  5. DecisionWe aim to reply within [X business days] of your conversation.

Questions

Do I need to know how AI models work?

No. You need to know how good decisions get made. We teach the task formats in a short onboarding.

Is this a job?

No. You work as an independent contractor on projects you choose.

Who are the clients?

AI labs and enterprise teams that build models. Client names are usually confidential, and we tell you what we can about each project before you accept it.

Can I do this while employed?

Many experts will. Check your employment agreement for limits on outside work. The work never involves your employer's information.

Why women and non-binary executives?

Because they are underrepresented in the record models learn from. Women hold 23 percent of C-suite roles across 57 economies.2

Where are you based, and can I apply from outside the US?

We are based in El Dorado Hills, California. The first cohort is US only. Leave your email and we will tell you when we open other countries.

Your judgment is what the models are missing.

Start your application

Sources

  1. Mercor homepage, checked November 6, 2026. Link to source 1 (opens in a new tab)
  2. World Economic Forum, Global Gender Gap Report 2026, September 16, 2026, using LinkedIn data. Link to source 2 (opens in a new tab)