Senior Full Stack Engineer

About the Company
A venture-backed B2B fintech of under 30 people, based in Texas. Its software runs credit and collections for companies that sell goods to other businesses: who gets payment terms, who gets a reminder, and which invoice each payment belongs to. The founders ran one of those businesses before they built this, and their customers already put billions of dollars of open invoices through it. The AI agents in the product do real work for those customers every day.
The Role
Own the systems that move a distributor's money.
Payments, credit decisions, collections, and the AI agents that now act inside them. You take a feature from the customer conversation to production, and you stay with it after launch.
You would join a small, flat team where everyone is an individual contributor, working in Ruby on Rails and PostgreSQL with React on the front end. Cursor and Claude Code are part of how the team ships, not a side experiment. You would build payment flows, connect the product to the customers' ERPs, and put LLM features into production with the evals and guardrails they need when real money is on the line. You would also sit on customer calls, because that is where the specs come from.
What is actually unsolved
Letting an agent act on money without a person checking every step. Today the product sends payment reminders in the credit manager's own voice, matches incoming checks and wires to open invoices, and posts the result into ERPs that were built decades ago. Each of those steps has to be right often enough, visible enough and easy enough to undo that a credit team stops double-checking it. Nobody has a playbook for that yet. This team is writing one, and you would be doing it with them.
Responsibilities
Build and run the payment processing systems and the financial APIs around them.
Connect the product to the ERPs that run distribution, NetSuite and Sage among them, in both directions.
Ship LLM-powered workflows to production: credit decisions, remittance matching, customer communication and dispute handling.
Build the evals, monitoring and guardrails that keep those features safe with real money on them.
Own the data pipelines that move millions of financial records in near real time.
Improve the collections automation, and measure whether it moved the customer's cash.
Join customer calls and read support threads, and turn what you hear into scoped, shippable work.
Requirements
You have shipped user-facing features to production end to end in the last two years: front end, backend, schema and tests.
Deep backend experience in Ruby on Rails, Django or Spring Boot, with PostgreSQL in production. You have designed schemas, run migrations on live tables and made slow queries fast.
Five or more years as a backend-leaning full-stack engineer, including time at a company under a hundred people that shipped new product fast.
You already use AI coding tools every day, and you know where they let you down.
Nice to Have
LLM or agent features in production, with evals, A/B tests or bandits behind them.
Payments, billing or fintech systems where being wrong costs money.
ERP integration work, on either side.
Ruby on Rails specifically.
US Central time zone, or an easy trip to Texas four times a year.
Culture
Everyone ships, everyone reviews, and people pair when they get stuck. The team spends more time on clear specs than on process, and whoever writes the spec usually ships it and answers for it afterwards. Four times a year everyone meets in Texas. One thing to know before you decide: the team calls this a startup grind, long hours and real intensity, in return for rebuilding the core systems of a company whose customers already depend on it. The role is open to US citizens and permanent residents only; it does not sponsor visas.

Work arrangement
No