Jeremy Zhang

One API to Rule 6,000 Systems

In 2020, Jeremy Zhang sat down to solve a problem that should have been simple: a lender needed payroll data to process Paycheck Protection Program loans. The answer was a mess of PDFs, emails, and manual entry - a process that hadn't meaningfully changed since the 1990s. Zhang and co-founder Ansel Parikh decided to fix the plumbing. What they found, once they started looking, was that the plumbing problem extended far beyond PPP loans. It was every HR software company, every fintech, every benefits provider - all rebuilding the same integrations to the same fragmented set of employment systems, over and over, forever.

That observation became Finch: a single API that connects to 200+ HR and payroll providers, translating the Tower of Babel that is American employment data into a unified, developer-friendly format. The pitch is straightforward enough to fit on a business card. The execution is the kind of thing that takes years of grinding through edge cases, data quirks, and provider negotiations that never end.

Five years in, Finch processes more than 5 million API calls every day. Its platform touches payroll for 18 million employees across tens of thousands of employers. The company raised a $40 million Series B in February 2023, co-led by General Catalyst and Menlo Ventures - the same investors who backed it from the seed round. In 2024, HackerNoon named Finch the Startup of the Year in San Francisco and ranked it third among all North American startups.

The Accidental Pivot

Zhang studied Electrical and Computer Engineering alongside Economics at Duke University - two disciplines that share an obsession with systems and signals. He wasn't just a student. During three semesters he founded an education startup, built the technology division of a campus enterprise, and wrote a portfolio management algorithm he called AntiCor. Backtested on NYSE data, it returned 293% annually on average. That number is either proof of something or a cautionary tale about overfitting - either way, it reveals someone who thinks naturally in terms of systems operating at scale.

After Duke, Zhang took a stint at Amazon Robotics - physical automation, before he discovered that software automation was a larger market - then joined Smartcar, a connected-car API company, as an early engineer. He spent two and a half years there, eventually leading product initiatives. The pattern was already clear: Zhang gravitates toward APIs, toward the infrastructure layer, toward the place where many systems need to talk to each other and currently cannot.

He met Ansel Parikh through On Deck's second cohort. They applied to Y Combinator together, got in for Summer 2020, and launched Finch in December of that year with a $3.5 million seed round from General Catalyst, Menlo Ventures, BoxGroup, Homebrew, and Clocktower Ventures.

Infrastructure Nobody Sees, Everyone Needs

The employment sector has a connectivity problem that makes the early internet look well-organized. About 6,000 HR and payroll systems operate in the U.S. They communicate via SFTP, spreadsheet uploads, and manual operations. Each one has its own data format, its own authentication flow, its own quirks. Every company building on top of employment data - benefits providers, insurance underwriters, retirement plan administrators, financial wellness apps - has to build integrations to whichever systems their customers use. Then maintain them. Then rebuild them when providers update their APIs.

Finch's answer is to build and maintain those integrations once, then expose them through a single standardized API. Developers connect to Finch; Finch connects to everything else. The coverage now reaches 88% of U.S. employers - which means most applications built on Finch work for most of their customers from day one, without additional integration work.

Finch offers SDKs in JavaScript, Python, Java, Kotlin, and Go. It handles OAuth authentication, webhook delivery, data encryption, and HR data compliance. The platform has become infrastructure for retirement plan integrations, benefits deductions, insurance underwriting, payroll synchronization, and employment verification - the invisible plumbing that financial and HR applications run on.

From $3.5M Seed to $58.5M Total

The funding story is notable for its consistency: the same investors who led the seed round - General Catalyst - co-led the Series B. Menlo Ventures, in from the Series A, co-led alongside them. Managing Director Matt Murphy joined Finch's board at Series B. The pattern suggests that the investors who saw the earliest version of the pitch kept increasing their conviction as the data came in.

Between the Series A and Series B, Finch grew revenue 12x and reached cash flow positivity - a combination that helps explain why the B round closed at $40 million during a period when many startups were struggling to raise at any valuation.

The Open Employment Ecosystem

Zhang frames Finch's mission as democratization - specifically, making employment data as accessible and interoperable as financial data became after Plaid, or web analytics after Segment. The analogy is precise: those companies didn't invent new categories, they unified fragmented ones. Before Plaid, connecting to a bank account meant SFTP or screen scraping. Before Segment, analytics integration meant adding a dozen different SDKs to your codebase. Before Finch, accessing payroll data meant PDF emails and manual operations.

The downstream consequence of what Finch is building - if it works at the scale Zhang envisions - is that a benefits startup can launch nationally without a dedicated integrations team. A retirement plan administrator can onboard a new employer in minutes rather than weeks. An insurance company can access employment verification data in real time rather than chasing down paystubs. The value proposition compounds as the network grows: more providers connected means more employers covered, which means more customers served, which means more reason for the next developer to build on Finch instead of rolling their own.

Zhang has described this as building an "open employment ecosystem" - a phrase that borrows from the open finance rhetoric of the post-Dodd-Frank era but applies it to employment data, which has largely escaped the regulatory attention that pushed banks toward API access. Whether the ecosystem remains open as Finch scales - or whether the company captures enough of the stack to become a tollbooth rather than a utility - is the interesting long-term question.

A Builder Who Thinks in Systems

Zhang's GitHub handle is jerzzhang - a compressed version of his name that mirrors exactly what Finch does to HR data. The GitHub itself is a window into how he thinks when he's not running a company: projects include an AI chief-of-staff framework built on Claude Code, a Gmail automation tool called AutoPrint, and early experiments with autonomous agents and AI-native workflows. The pattern is consistent - automation of repetitive systems, applied at whatever layer presents itself.

He was also one of 14 Duke undergraduates selected for True Ventures' entrepreneurship and venture capital fellowship - a program designed to give exceptional students a view of how early-stage investing works before they graduate. The experience likely sharpened an intuition for what makes a company fundable, which shows up in how clearly Zhang frames Finch's market position and competitive differentiation in interviews and press materials.

Zhang met co-founder Ansel Parikh through On Deck's second cohort (ODF2). Parikh came from the investment side - Kleiner Perkins and Bond - bringing a finance-and-operations counterbalance to Zhang's engineering and product background. The pairing is characteristic of successful technical founders who find a co-founder who complements rather than mirrors them.

The Path to Finch