Finance Prompt Engineer

Finance Prompt Engineers are the most distinctive role at Equiforte — and among the most critical. You are not a traditional engineer, and you are not a traditional finance professional. You are both: a domain expert with deep, hands-on experience in private capital finance who has developed the instinct for how AI systems can and cannot be trusted with financial data.

In this role, you sit with clients. You understand their reporting workflows in the same language their CFOs use. You know what a correctly structured waterfall looks like, what a compliant LP letter requires, and where AI outputs need human judgment before they ship. You translate that understanding into the prompts, validation logic, and workflow configurations that make Equiforte's platform produce results that clients trust absolutely.

This is a forward-deployed role. You will be embedded with client firms — attending their quarterly close cycles, sitting in on their reporting reviews, and acting as Equiforte's on-the-ground expert as they adopt AI-native workflows. Your feedback loops directly into product: every pattern you encounter in the field becomes an improvement to the platform.

What you'll do

• Work directly with CFOs, Controllers, and fund finance teams at client firms to understand their specific reporting workflows, data structures, and quality standards
• Design, test, and refine the prompts, templates, and extraction logic that drive Equiforte's AI outputs for each client's use cases — LP reporting, capital call processing, valuation documentation, audit packages
• Validate AI-generated financial outputs against source data and LP agreements before client review, bringing the same rigor to AI outputs that a senior associate would bring to a spreadsheet model
• Configure client-specific workflow logic — allocation rules, side letter provisions, waterfall structures, LP formatting preferences — within the Equiforte platform
• Train client finance teams on how to work effectively with AI-generated outputs — what to trust, what to verify, and how to identify exceptions
• Build reusable prompt libraries, validation frameworks, and configuration templates that the broader team can deploy across similar client profiles
• Feed structured product feedback to engineering based on patterns, failure modes, and client-specific requirements encountered in the field

This role reports to the Head of Client Delivery and works closely with the Engineering, ML Engineering, and Product teams.

• You have 5+ years of hands-on private capital finance experience — fund accounting, LP reporting, capital call processing, valuation, or quarterly close at a private equity, private credit, or infrastructure firm, or at a fund administrator serving those firms
• You are deeply familiar with the mechanics of institutional LP reporting: ILPA standards, side letter provisions, carried interest calculations, and the documentation requirements of sophisticated LPs
• You have started using AI tools seriously and have developed strong intuitions about where they produce reliable outputs and where they fail — especially in financial contexts
• You are technically curious: you do not need to write code, but you are comfortable working inside software systems, configuring logic, and debugging output quality issues
• You are effective in client-facing environments at the CFO and Controller level — you communicate with precision and can hold your own in conversations about financial methodology
• You thrive in ambiguous, early-stage environments where the playbook is being written in real time
• You want to be a builder, not just an implementer — you see your field experience as a resource for making the product better

Please be prepared to address the following when you apply:

• Describe a quarterly close or LP reporting process you have owned or been deeply involved in. What were the most error-prone steps, and how were they managed?
• Have you used AI tools — including LLMs or document extraction tools — in a financial context? What worked, what did not, and how did you validate the outputs?
• What aspect of private capital finance do you think is hardest to automate reliably, and why?