Software Engineer (Full Stack)

Software Engineer (Full Stack), New York

Creeda is an AI-powered franchise accounting platform. We sell live financial visibility to franchises so they can run their businesses better, and automation to franchise-focused bookkeeping firms so they can reduce costs and increase efficiency.
There are 850,000 franchises in the US, constituting a $921B industry. Creeda is growing very fast, so we are looking for the brightest engineers to help us execute.

The role

You'll start by building. We need a strong generalist who can work anywhere in the stack, front end to back end to data, and make the platform better wherever it needs it most. Some weeks that's shipping a feature a customer is waiting on, some weeks it's fixing an integration that breaks a bookkeeper's close, some weeks it's making onboarding faster so a firm can bring over hundreds of locations instead of fifty. You take a problem and own it end to end.
As we sign larger firms, the job will become more of a FDE role. You'll fly out, sit with their accountants, watch how they actually work, and build automations on site that go straight into production. Their bookkeepers describe what they need, you build it that week, and every firm serving that brand gets it. Half engineer, half the person our customers call when they want something built.
This is one of the highest-leverage seats at Creeda. You'll see the customer, write the code, and ship the fix in the same day.

Who this is not for

• Engineers who want clean specs and a well-defined surface. There is no such thing here yet.


• Engineers who do not want to talk to customers. Roughly a third of this job is people.


• Engineers who want to work only on the interesting parts. The data cleaning is not a phase we are going to grow out of; it is the product.

What we are looking for

Required

• Strong backend engineering in TypeScript or Python. We run both (TypeScript on Node and Next.js for the product, workers and importers; Python for the classifier and ML training) on Postgres. You need to be strong in one and willing to work in the other.


• Real experience with data that was not made for you: documents meant for humans (PDFs, spreadsheets a bookkeeper has already edited), vendor exports, third-party APIs that contradict themselves.


• Enough rigor to work on financial data, where being approximately right is being wrong. You should be able to explain why a balance check passing proves the arithmetic and not the entry.


• Judgment about code**.** We build with AI coding agents, and they are fast enough that the bottleneck has moved to review. You need software design fundamentals strong enough to look at generated code and say precisely why it is wrong. You should be good at the other end of it too, setting constraints up front so the agent produces something reviewable in the first place.


• Clear written communication. Every area of the codebase carries its own orientation doc, read by the next engineer and by the AI coding agents we build with, and we treat that documentation as an asset.


• Willingness to travel periodically once phase two begins, several multi-day trips per quarter

Helpful, not required

• Putting ML or LLM output into production behind hard constraints, where the model is allowed to be wrong but the system is not


• Double-entry bookkeeping. You know what a journal entry is and why it has to balance.


• Accounting, fintech, POS, payroll or ERP background


• QuickBooks Online's API, Plaid, or any general ledger system


• Having worked somewhere small enough that titles did not decide who did what
  We do not require a CS degree, and we do not care where you worked before. We care that you have built something that had to be right.

Details

• Location: New York City


• Visa: We can accept CPT/OPT but unfortunately cannot sponsor H1B

Hiring Process

1. Please send your resume and a brief self-introduction to [email protected]


2. 30 minutes with the CTO on what you have built


3. A working session on a real piece of our data, not a puzzle. We will show you a payout statement or a payroll register the way it actually arrives, and work through turning it into an entry that balances. Part of it will be reviewing code an agent wrote for the same task, and telling us what you would not ship.


4. A conversation with the CEO about the business, because you will be in front of customers


5. References
   We aim to go from first call to offer in under two weeks.