Finance Expert -- AI Training & Evaluation

We’re looking for experienced finance professionals to help train and evaluate frontier AI agents on computer-use finance tasks. You should have 3+ years of hands-on finance experience in areas such as FP&A, corporate finance, strategic finance, financial analysis, controllership, treasury, investment analysis, or related work.

\n

\n

About Plato

\n

\n

Plato is building the environments and evaluations needed to train the next generation of AI agents.

\n

Our focus is on computer-use agents: AI systems that can work across the same software, files, and digital workflows that professionals use every day. We want to understand what finance professionals actually do on their computers, what great work looks like, and where today’s best agents still fall short.

\n

\n

Your expertise will directly help shape the real-world tasks and feedback used to train and evaluate frontier AI systems.

\n

\n

What you'll do

\n

\n

You’ll help us understand and recreate real finance workflows performed on a computer. This may include:

\n

• \n
• Walking us through tasks you perform across tools such as Excel, Google Sheets, ERP systems, browsers, email, and internal software
• \n
• Recreating realistic finance workflows involving multiple applications, files, and sources of information
• \n
• Testing AI agents on these workflows on our proprietary platform
• \n
• Identifying mistakes, edge cases, and failure modes
• \n
• Explaining what a strong professional-quality outcome looks like as reward criteria or rubrics
• \n

\n

\n

Examples might include building or updating financial models, investigating variances, reconciling data across systems, preparing reports, reviewing financial information, or completing multi-step workflows across spreadsheets and business software.

\n

\n

Who we're looking for

\n

\n

You may be a strong fit if:

\n

• \n
• You have 3+ years of professional finance experience
• \n
• You have recent, hands-on experience doing finance work yourself
• \n
• A meaningful part of your work happens on a computer across spreadsheets, browsers, email, financial systems, or other business software
• \n
• You can give concrete examples of real multi-step workflows from your work
• \n
• You have heavily used tools such as Claude, ChatGPT, Copilot, or Gemini in your professional work and/or have previous AI training or evaluation experience with companies such as Mercor, Handshake, Scale, Surge, or Turing
• \n
• You can clearly explain what makes financial work good, bad, incomplete, or unrealistic
• \n
• You have enough flexibility to contribute consistently alongside your existing work
• \n

\n

\n

This is a flexible, ongoing paid role designed to be done alongside your existing work. $75–125+/hour, depending on experience.