Julius Expert Earners

Train Financial Superintelligence

Join the expert network of finance professionals to create real-world tasks, reference solutions, and evaluations that help improve advanced AI.

  • $100+ per task
  • ~$200/hour on average
  • Remote, flexible schedule
Apply now

Work with experts from

  • Harvard Business School
  • Boston Consulting Group
  • Princeton University
  • NVIDIA
  • Centerview Partners

Why train AI with Julius

Julius is an AI lab in San Francisco building AI tools used by over 2,000,000 people. Expert Earners pays finance professionals to show AI what good work looks like.

Earn $100+ per task

Every task pays a $100 baseline, with reward milestones on top. Experts average about $200 an hour.

Set your own schedule

Work when and where you want, starting at 10 hours a week. It fits around a full-time job or MBA classes.

Build what models learn from

Your tasks and reference solutions become the examples AI models are trained and graded on.

Steady work

New projects keep coming, and you decide how much to take on.

Free AI tools

We provide licenses to the AI services you use on projects at no cost to you.

Work directly with an AI lab

Expert Earners is run by Julius, an AI lab in San Francisco, not a staffing marketplace.

About the role

Design the finance problems AI learns from

You write the hard, realistic tasks an analyst or associate actually gets, solve them yourself, and judge how AI models handle them.

What you'll do

  • Write realistic finance tasks, like building a DCF, linking a three-statement model, or running a sensitivity analysis
  • Create reference solutions with the formulas, assumptions, and answers an expert would deliver
  • Write grading rubrics and evaluate AI responses for errors in the numbers and the logic
  • Find the financial workflows where today’s AI models break

What we look for

  • 2 to 5 years of experience in investment banking, private equity, equity research, corporate finance, FP&A, accounting, or finance-heavy consulting
  • MBA students and recent graduates are encouraged to apply
  • Credentials like the CFA or CPA are a plus
  • Advanced Excel or Google Sheets skills, and fluency in the income statement, balance sheet, and statement of cash flows
  • The attention to detail to catch a broken link in someone else’s model
  • Curiosity about AI. No technical background required

Example task · Valuation

$100+

DCF sensitivity for a SaaS company

From three years of income statements, balance sheets, and cash flow statements, build a five-year unlevered DCF. Then show how enterprise value moves with WACC and terminal growth.

Enterprise value ($M)
WACC \ Growth2.0%2.5%3.0%
8.0%9321,0021,087
9.0%794843901
10.0%691727768

What you submit

  • Task prompt
  • Reference solution
  • Grading rubric

The Hiring Process

Four steps, each a preview of the actual work.

  1. 1

    Resume review

    Within 3 business days

    Tell us about your finance experience and your biggest wins.

  2. 2

    Technical assessment

    60 min

    Show how you reason through problems from your area of finance.

  3. 3

    Screening call

    30 min

    Go deeper on your experience and how you use AI.

  4. 4

    Mock task build

    3 business days

    Build a task that exposes a model weakness. We provide the inputs and never use your output.

Pass all four steps and you're ready to start earning $100+ per task.

Apply now

Apply to become an Expert Earner

Tell us about your background. It takes about five minutes, and we review applications within three business days.

  • $100+ per task
  • ~$200/hour on average
  • Remote, flexible schedule

By applying, you agree to our Privacy Policy.

Looking for DataAnnotation, Outlier, or Mercor alternatives?

Expert Earners is built for finance professionals who want paid work that uses what they know.

Built for finance

Projects center on finance, accounting, and spreadsheet work instead of general annotation.

Pay that reflects expertise

Tasks start at $100 with reward milestones on top, and experts average about $200 an hour.

Expert-level work

You design the task and the reference answer yourself instead of only rating AI outputs.

Run by Julius

Work directly with the AI lab behind Julius instead of a third-party marketplace.

Frequently asked questions

What is an AI training job?
An AI training job pays you to create the examples AI models learn from. As an AI trainer with Julius Expert Earners, you write realistic tasks from your field, build a reference solution, and evaluate how well AI models handle them.
What kind of tasks will I create?
Tasks look like real assignments: build a DCF from historical financials, link the income statement, balance sheet, and statement of cash flows, run a sensitivity analysis on WACC and terminal growth, or find the error in a broken model. You also write the reference solution and the rubric used to grade AI answers.
How much do AI training jobs at Julius pay?
Every task pays a $100 baseline, and you can earn reward milestones on top of that. Tasks take anywhere from 15 minutes to a couple of hours, and experts average about $200 an hour.
How do I get paid?
Experts work as independent contractors and are paid by direct deposit.
Who should apply?
Finance professionals with 2 to 5 years of experience in investment banking, private equity, equity research, corporate finance, FP&A, or accounting. MBA students and recent graduates are encouraged to apply, and credentials like the CFA or CPA help. Consultants are a fit when their work is heavily finance-focused.
Do I need AI or coding experience?
No. We are looking for finance expertise. If you can build a three-statement model, run a DCF, or set up a sensitivity table, you have the skills that matter.
Do I need to pay for AI tools?
No. We provide licenses to the AI services you need for projects, free of charge.
Is there a minimum number of hours?
Yes. We ask for at least 10 hours a week. Beyond that there is no fixed schedule: the work is async, so you can work whenever suits you, including evenings and weekends, and take on more as projects allow.
Is the work steady?
Yes. Engagements are ongoing rather than fixed-length, and experts in good standing are rolled onto new projects as they open up.
Can I do this alongside a full-time job?
Yes. The work is async, so it fits around a full-time job or an MBA program. You are responsible for following your employer’s policies, including any rules on outside work and confidential information.
Is there an NDA?
Yes. Every expert signs a nondisclosure agreement before starting project work. It covers the tasks you create, the AI outputs you review, and any materials we share, so project details stay confidential and can’t be posted publicly or shared outside Expert Earners. You can keep your full-time job while working with us, but never bring your employer’s confidential information into a task.
Is this a remote job?
Yes. All Expert Earners projects are fully remote.
What happens after I apply?
We review applications within three business days. Next comes a 60-minute technical assessment, a 30-minute screening call, and a mock task build you complete over three business days. Ready to start? Apply with the form on this page.
How does this compare to RLHF and AI evaluator jobs?
RLHF (reinforcement learning from human feedback) and AI evaluator jobs ask people to judge model outputs. Expert Earners includes evaluation, and you also design the task and the expert answer, which takes deeper domain knowledge.
How is Expert Earners different from DataAnnotation, Outlier, or Mercor?
DataAnnotation, Outlier, and Mercor recruit for many kinds of AI training work. Expert Earners focuses on finance expertise, pays $100+ per task, and is run directly by Julius.