Founding Engineer

About the company

Our client is a seed-stage, venture-backed AI startup in New York automating the paperwork that keeps patients waiting for specialty treatments. Its platform connects providers, payers and pharmacies, and uses AI to handle prior authorization, benefits verification and enrollment in manufacturer assistance programs — the steps that normally add days or weeks before a prescription is actually filled. Founded in 2024 and backed by a $7M Seed, it already works with hundreds of providers across 15 states, processes more than 2,000 prescriptions a month, and has been roughly doubling its doctor and patient footprint every month. The team is four people today, which is why this is employee number five.

The role

You would join as a founding engineer and employee number five, reporting straight to the co-founder and CTO on a small, deeply technical team. The work is integration-heavy and agentic: building into electronic medical records on the provider side, pharmacy management systems, and payer systems for electronic prior authorization, then architecting the high-throughput Python workflows — a hybrid of browser automation and direct APIs — that keep those integrations running at real volume. You would also build the core product used by clinics, patients and the internal operations team, which means genuine time with clients and clinical support staff rather than only with the codebase. The client is hiring two engineers here: one strongly backend, and one backend-leaning full-stack with real React. The pattern of the job is to prototype fast, then turn what works into production-grade systems.

What you'll do

• Build agentic integrations into provider electronic medical records, pharmacy management systems and payer electronic prior-authorization systems
• Architect and scale high-throughput Python workflows that combine browser automation with direct API integrations
• Build the core product used by clinics, patients and the internal operations team, working closely with clients and clinical support staff
• Push AI further across the medication workflow — more accuracy, and more of the complex steps automated end to end
• Prototype quickly, then harden what works into production systems the team can rely on
• Help build the platform and integrations that take the client from hundreds of doctors to more than a thousand

Nice to have

• Has built a product or system from scratch, rather than only incremental improvements to an existing one
• Has shipped front-end work in React alongside backend work
• Background in healthcare, health tech or government tech
• A computer-science degree from a top-30 program
• Has built integrations into third-party systems of record, browser automation, or high-throughput Python pipelines
• Has built LLM or agent features in production

Tech stack

Python across the backend with React on the front end, PostgreSQL and Redis for data, and a mix of Selenium-driven browser automation and direct API integrations for the systems of record this product has to talk to. LLMs and agentic workflows sit at the centre of the product rather than at its edges. The work leans backend on both hires: roughly 80% backend or more for one, and about two thirds backend to one third React for the other.

Compensation & logistics

• $160k–$220k base plus 0.5–1.5% equity
• Hybrid in New York — three or more days a week in the Flatiron office, and the team tells candidates to expect three to four
• Open to candidates who need no sponsorship and to simple visa transfers such as an H-1B transfer or OPT; the client will not file a new, first-time visa sponsorship
• Two hires on this role, both as founding engineers — one strongly backend, one backend-leaning full-stack
• Interview process: a 30-minute founder screen with the CTO, a two to three hour take-home built around a real integration task, a one-hour live technical screen covering system design or full-stack coding in React and Python, then a few hours on-site with the team — with an optional paid work trial for candidates who have fewer reference signals
• Full-time