Building the next generation of Finance Engineers

A Finance Engineer is a finance professional who builds the systems, agents, and automations that make every part of the finance function faster. They started as accountants, controllers, or FP&A analysts and taught themselves to use AI tools to architect workflows that used to require engineering.

They are the person the rest of the finance team goes to when they need something that doesn't exist yet. They see a manual close task and think: this should run itself. The role is emerging now the way Marketing Engineer, GTM Engineer, and Data Engineer did in previous eras.

Rillet is building the first certification for Finance Engineers. To qualify, you need to have attended at least 3 webinars or 1 live vibe coding session and built at least one AI-powered workflow a real finance team uses in production.

Submit your build through the certification waitlist. The first cohort will define the standard for the role. Certification launches soon.

FinOps analysts and systems accountants keep the infrastructure running: clean data, working integrations, reliable processes. If nothing breaks, they have done their job.

Finance Engineers build on that foundation to invent new capabilities. They are measured on the same KPIs as any other finance role (close speed, forecast accuracy, audit readiness, cycle time) but their output is systems, not spreadsheets. Most finance teams already have ops. The Finance Engineer is what's missing.

Vibe coding for finance is the practice of using AI tools like Claude, Cursor, Lovable, and Replit to build functional finance workflows, dashboards, and internal tools by describing what you want in plain English instead of writing code from scratch.

The term was popularized by Andrej Karpathy in early 2025. For finance teams, it means a controller can build a cash forecast dashboard, an AP automation tool, or a custom close checklist in an afternoon, without waiting on engineering.

No. Vibe coding is specifically designed for non-engineers. You describe the outcome you want in natural language (for example, "build me a dashboard that pulls my AR aging from Rillet and flags invoices over 60 days past due"), and the AI generates the underlying code, UI, and logic.

The skill being developed is prompt clarity and system thinking, not syntax. If you can write a clear spec in a Notion doc, you can vibe code.

The most common tools finance teams use are:

Claude for reasoning, analysis, and long-context work like close reviews and variance explanations. Claude Code and Cursor for building scripts, automations, and internal tools. Lovable and Replit for building full apps and dashboards from a prompt. Claude in Excel for spreadsheet-native AI. Claude Cowork for connecting AI to your files, apps, and documents.

MCP servers like the Rillet MCP connect these tools directly to your finance data sources. For a deeper walkthrough, see the Finance Guide to Claude & MCP.

MCP stands for Model Context Protocol. It is an open standard that lets AI tools like Claude connect directly to your systems, including your ERP, your spreadsheets, and your internal databases.

For finance teams, MCP is what turns generic AI into an accounting-aware assistant. Instead of copy-pasting a trial balance into Claude, you can give Claude live access to your GL through an MCP server and ask real questions like "which accounts moved more than 20 percent this month and why."

The Finance Guide to Claude & MCP walks through setup, use cases, and guardrails.

It depends on the tool, the deployment, and how you configure access. Enterprise-grade AI tools (Claude for Work, ChatGPT Enterprise, Cursor Business) offer data residency, no training on inputs, and SOC 2 coverage.

For sensitive production data, finance teams should use sandboxed MCP connections with read-only access, avoid pasting raw customer PII into consumer AI tools, and keep any generated automation under human review until it's proven in parallel with the existing process.

Real examples from the Rillet workshop series include:

A 13-week cash forecast dashboard. A procurement intake tool that routes requests through approvals. A close checklist that auto-updates based on journal entry status. A headcount plan reconciler that compares HRIS data to GL payroll. A board memo drafter that pulls from the GL and flags variances. A deferred revenue waterfall validator. A contract-to-cash tracer that follows a single deal from Salesforce opportunity to recognized revenue.

None of these required a software engineer.

Most of the tools demoed in the workshop are built in 30 to 60 minutes, which is why the sessions run exactly 60 minutes.

The first version is always rough. The second iteration, informed by actually using it, is where the value shows up. Finance teams typically spend one session building, then one to two more sessions refining before rolling a tool out to the team.

Controllers, VPs of Finance, CFOs, senior accountants, and finance ops leaders at growing companies. The sessions assume finance fluency and zero engineering background.

If you have ever thought "I wish we had a tool for this" and been told engineering could get to it next quarter, this is for you.

Sessions run every Friday for 60 minutes, hosted by Stephen Hedlund, Head of Finance at Rillet. Registration is free and all sessions are recorded and posted on this page.

Register for the next session, or scroll up to watch any replay.