The New Finance Role Taking Over Silicon Valley

Back in January, I sat with my co-founder Alexandre Labreche at a16z's SF office and talked about something that had been on my mind for a while: Finance people have always been builders at heart.

Look at the proof.

The sprawling Excel models, the macros held together with hope (and one very brave VLOOKUP), the data systems that finance teams designed, wired, and quietly maintained for years, while everyone else assumed they just "did the numbers."

But on that day in January, I had noticed a shift. For the first time, AI was handing finance the tools to build more cool shit. No longer was building new systems gated behind a request for engineering capacity or an IT portal. I was (understandably) very, very excited.

At first, this new domain didn't have a name. Just a handful of finance leaders who saw the opening and doubled down on getting stuff done. They experimented with new tech. They built workflows and tracked what worked, and what didn't. They found where the shiny new tools break and figured out what actually needed building to pull finance into the future.

Slowly but surely, companies saw the incredible value these finance teams unlocked, and over the next six months, it quietly became the hottest role in tech: The Finance EngineerGlossaryFinance engineerA finance engineer is a finance professional who combines accounting or FP&A judgment with data, automation, systems design, and AI workflow skills. For finance engineer, the useful boundary is the data, tools, approvals, human review, evaluation standard, and decision the system may influence.Open full definition →.

So let's talk about what it actually means, why it showed up now, and how to build the skill before the market forces your hand.

So what is a Finance Engineer?

Start with the simplest version. A Finance Engineer is a finance pro who has added a build-first, technical layer on top of everything they already know.

The finance foundation stays. Accounting standards, business context, commercial judgment... all of it still matters. What changes is the output. Controllers still close the books, but now they automate the close. FP&AGlossaryFinancial planning and analysisFinancial planning and analysis is the finance function responsible for budgeting, forecasting, analysis, and decision support. For financial planning and analysis, the useful boundary is the driver, assumption, source data, owner, time period, scenario logic, and decision the model is meant to support.Open full definition → leads still forecast, but now they build the dashboard instead of hand-feeding it every Monday or shipping the data off to a BI team. Heads of Finance still run the function, but now they build the systems that keep it moving smoothly.

To sum it up: They figure out their own requirements AND build the solution.

This is the heart of AI-native financeGlossaryAI-native financeAI-native finance is finance functions and tools built around AI and automation from the ground up rather than added onto legacy processes. For AI-native finance, the useful boundary is the data, tools, approvals, human review, evaluation standard, and decision the system may influence.Open full definition →. The work used to flow one way: finance describes what it needs, then engineering builds it (eventually).

The Finance Engineer collapses that gap. The person who understands the number is the same person who builds the thing that produces it.

What does a Finance Engineer actually do?

The day-to-day comes down to five skills. They line up neatly with what the best finance teams are already doing.

LLMGlossaryLarge language modelA large language model is a neural network trained on large text datasets to understand, summarize, and generate language. For large language model, the useful boundary is the data it uses, the tools it can call, the approvals it needs, the review standard, and the finance decision it may influence before the output is trusted or automated.Open full definition → literacy. Knowing how to put ChatGPT, CopilotGlossaryAI copilotAI copilot is an assistant that augments a human's work in context. For AI copilot, the useful boundary is the data it uses, the tools it can call, the approvals it needs, the review standard, and the finance decision it may influence before the output is trusted or automated.Open full definition →, or Claude to real work. A ControllerGlossaryControllershipControllership is the function responsible for accounting accuracy, the close, internal controls, and compliance. For controllership, the important details are the accounting period, source evidence, reviewer, materiality threshold, and control purpose that make the treatment auditable during close, reporting, and later review.Open full definition → can draft variance commentaryGlossaryVariance commentaryVariance commentary is the written explanation accompanying variances in management or board reporting. For variance commentary, the useful boundary is the driver, assumption, source data, owner, time period, scenario logic, and decision the model is meant to support. For variance commentary, the practical standard is making the driver, assumption, and owner visible when actuals change.Open full definition → on the monthly closeGlossaryMonth-end closeMonth-end close is the practice of recurring process of finalizing the books at period end so the financials can be reported. For month-end close, the important details are the accounting period, source evidence, reviewer, materiality threshold, and control purpose that make the treatment auditable during close, reporting, and later review.Open full definition →, grounded in the actual numbers, in minutes instead of hours. The trick is not just promptingGlossaryPrompt engineeringPrompt engineering is the practice of designing instructions, examples, constraints, and context so an AI model produces more useful outputs. For prompt engineering, the useful boundary is the data it uses, the tools it can call, the approvals it needs, the review standard, and the finance decision it may influence before the output is trusted or automated.Open full definition →. It is validating the output against the source so you trust what comes back.

Finance automationGlossaryFinance automationFinance automation is the practice of using software and AI to automate finance and accounting tasks such as data consolidation, reporting, and analysis. For finance automation, the useful boundary is the data, tools, approvals, human review, evaluation standard, and decision the system may influence.Open full definition →. Spotting friction and removing it. An FP&A lead refreshes the dashboard daily instead of weekly because the data pull runs itself. A finance ops lead turns reconciliationGlossaryAccount reconciliationAccount reconciliation is the practice of matching an account balance to supporting source data and resolving any differences. For account reconciliation, the important details are the accounting period, source evidence, reviewer, materiality threshold, and control purpose that make the treatment auditable during close, reporting, and later review.Open full definition → from a Monday morning slog into a quiet background process that only pings when something looks off.

Data literacy and governance. Reading data structures, working across messy and clean sources alike, and catching when an input is incomplete before it wrecks a report. This is what lets you dig into the data modelGlossaryData modelData model is an abstract representation of how data elements relate and are organized within a system, spanning conceptual, logical, and physical levels. For data model, a useful definition states an abstract representation of how data elements relate and are organized within a system, spanning conceptual, logical, and, who owns it, and which decision it supports.Open full definition → yourself rather than wait three days for someone else to answer it.

Systems integration. Understanding how data moves through the ERP, and being able to connect tools with basic SQL or API know-how. A Head of Finance who can link the ERP to a planning tool, without filing an engineering ticket, is never blocked by the plumbing.

Context and knowledge management. This is the one that makes the rest work at enterprise scale. AI is only as good as the context you feed it. So the real unlock is a persistent, clean, readable record of how your organization actually works, sometimes shaped into a knowledge graphGlossaryKnowledge graphKnowledge graph is a graph-based representation of real-world entities and the relationships between them, used to connect data across sources and support querying and reasoning. For knowledge graph, the useful boundary is the data, tools, approvals, human review, evaluation standard, and decision the system may influence.Open full definition → or an ontologyGlossaryOntologyOntology is a formal model of the concepts, entities, and relationships in a domain, including hierarchies and constraints that support inference, distinct from the analytic semantic layer. For ontology, a useful definition states a formal model of the concepts, entities, and relationships in a domain, including hierarchies and constraints that support, who owns it, and which decision it supports.Open full definition →. For a finance team that wants to use AI for genuine reasoning or analysis, that context layer is not a nice-to-have. GroundingGlossaryGroundingGrounding is anchoring model outputs in verified, retrievable source data. For grounding, the useful boundary is the data it uses, the tools it can call, the approvals it needs, the review standard, and the finance decision it may influence before the output is trusted or automated.Open full definition → the work in your own specifics, your definitions, your structure, your history, is what turns a generic model into a useful one. And when that context lives in an asset your function owns, rather than locked inside someone else's tool, it only grows more valuable over time.

Notice the pattern. None of this asks finance to become software engineers. It asks them to operate at the system level, using tools they can actually pick up.

The data backs the priority order. In CFO Connect's research, 56% of finance pros rank workflow automation as the single most important AI skill for their career over the next two years. Data literacy lands at 21%. Cross-functional work with AI sits at 14%. Automation is the front door.

Why now?

Fair question. Finance teams have always been scrappy, so why did this role show up in 2026 and not five years ago?

A few things lined up at once.

First, the tools finally got good. Large language models crossed the line from a neat demo to a tool you can lean on. They read unstructured data, draft narratives, and check their work against source documents. That was science fiction a couple of years back.

Second, the wiring got simple. Connecting an AI tool to live finance data used to mean a project and a budget. Now protocols like MCPGlossaryModel Context Protocol (MCP)Model context protocol (MCP) is an open standard introduced by Anthropic in November 2024 that standardizes how AI applications connect to external tools, data, and systems through a client-server architecture. For model context protocol (MCP), the useful boundary is the data, tools, approvals, human review, evaluation standard, and decision the system may influence.Open full definition → let a finance lead plug a model straight into the spend data with little more than a single address. This is what people mean by agentic financeGlossaryAgentic financeAgentic finance is the practice of using autonomous AI agents to perceive, plan, and execute multi-step finance tasks with limited human intervention. For agentic finance, the useful boundary is the data, tools, approvals, human review, evaluation standard, and decision the system may influence.Open full definition →: AI that does not just answer questions, but takes action across your systems and runs multi-step work on your behalf.

Third, the proof showed up. Real CFOs started demoing real automations to rooms full of peers. Once one person shows that a three-hour report can run in eight minutes, the room does not forget it.

Fourth, the market is moving. The job title is appearing on real postings with real pay attached.

Put those together and you get the moment we are in. The capability is here, the friction is gone, the proof is everywhere, and the hiring has started.

What this looks like in real life

Skip the theory for a second. Picture a CFO whose week used to start the same painful way: open Excel, open the ERP, export, open the spend tool, export again, then VLOOKUP and pivot for hours until a trustworthy number finally appeared. By then the number was already a week old.

Her fix was not a new data team or another ticket in the queue. She connected an AI assistant straight to the company's live spend data herself. Three core reports that used to eat hours each, travel compliance, OPEX review, and AP aging, now run in under eight minutes apiece. She asks a plain-English question and gets an answer she can trust.

That is the four skills working together at the very top of the org. LLM literacy, finance automation, data fluency, and systems integration, all in one move. She built the tool. She did not wait for it.

Who is hiring for this?

The clearest public signal comes out of the US. Spring Health, a Series E mental healthcare company valued around $3.3 billion, is openly hiring a Senior Finance Engineer.

The role reports into the VP of FP&A and the Chief of Staff to the CFO. It sits inside finance, not off in a separate engineering org. The ask is roughly four-plus years deploying automation or AI in a finance setting, paired with four-plus years in FP&A, strategic financeGlossaryStrategic financeStrategic finance is the practice of forward-looking finance work focused on driving decisions, capital allocation, and growth. For strategic finance, the useful boundary is the driver, assumption, source data, owner, time period, scenario logic, and decision the model is meant to support.Open full definition →, or accounting. The posted base range runs $159,100 to $179,000, plus equity and benefits.

It's worth noting that this comes in at nearly 23% higher than their other open Sr Finance roles.

This is NOT a software engineer being asked to learn finance. It is a finance person being asked to operate at the system level.

Job title, or skill layer?

In my opinion? Both. But the ratio is worth looking at.

In 2026, this is overwhelmingly a skill layer. Most people building these abilities are not changing their title at all. They are changing what they can do. The Controller automating the close is a Finance Engineer, whether or not the job description says so. So is the FP&A lead who builds the dashboard and automates the refresh.

A small but growing group of companies are turning that layer into a dedicated senior role with its own reporting line and pay band. That makes sense once the automation work gets big enough to need a specialist instead of a shared habit.

For most finance leaders, this means treat the Finance Engineer as a capability to build inside the team you already have, not a brand-new headcount you need to go fight for.

Build the muscles first. The title follows the work. And if the role does firm up over the next couple of years, which the early signs suggest, the teams that started early get to promote from within instead of scrambling to hire.

How to start this quarter

Three moves. None of them need a budget approval.

1. Automate one workflow, end to end, this month. Pick something that eats manual time right now: the close, variance commentary, the dashboard refresh, or reconciliations. Then automate the whole thing with tools your team already has. This is build-and-measure, not a procurement project. The lesson lives in shipping something and iterating, not in comparing vendor decks.
2. Audit your team's AI fluency, honestly. More than half of finance pros now see workflow automation as the top career skill for the next two years. If your team is not actively building that muscle, you are already behind your peer set. That gap compounds fast at the current pace.
3. Update every finance hiring brief in 2026. For FP&A leads, controllers, and ops roles, write automation literacy and AI fluency into the requirements, not the nice-to-haves. The expectations are shifting at every level, not just the senior one. AI for CFOs is no longer a special project. It is the baseline.

The takeaway

Finance never lacked builders. It lacked tools that matched the drive. That part is solved.

The Finance Engineer skill layer is already forming inside your team. Someone is automating something this week without being asked. The only real question is whether you build that capability on purpose, or notice it after the fact when a competitor's job posting reminds you.

Build it on purpose. Finance has been waiting a long time for this.

— Vanessa

About the author

Vanessa Galarneau

CFO & COO

LinkedIn →