One thing up front, because it's the honest version: if you go looking for this company online you won't find much. Small team, no marketing function, website that undersells them. That's not a warning sign — it's a thirteen-person company that has spent its attention on customers. We've done the diligence and we'll walk you through everything: funding, investors, revenue trajectory, the interview process, who you'd be working with. Ask us anything.

What they do

Specialty medications are among the most expensive drugs in the world, and the barrier isn't the medicine — it's the paperwork. Prior authorizations, benefits verification, pharmacy routing, assistance programs. Specialty practices lose dozens of hours a week to it and patients wait.

Our client's AI agents deploy into a provider's own systems and run the whole path from provider to patient to pharmacy. The product is free to providers;

Their framing, which we liked: they're not building an AI product, they're making doctors' lives better and happen to use AI to do it.

The role

Two levels, same job. Where you land depends on what you've built.

• Provider-facing product across Python, React/TypeScript and Postgres.

• AI automation for prior authorization and benefits verification, including agentic work against third-party systems that will never expose an API.

• Internal AI tooling — local model hosting, company-specific harnesses, PHI-compliant pipelines. PHI and third-party models don't mix, which is what makes it interesting.

• Owning projects from spotting the problem through shipping it and monitoring what happened.

• Direct work with both founders. Four or five engineers today, potentially double that in six months.

What they're strict about

• Startup experience. Early-stage or high-growth. Founder and founding-engineer backgrounds are among their favourites, especially zero-to-real-revenue.

• End-to-end ownership. A problem you found, built, and then monitored. Not a ticket you closed.

• Python without an AI assistant carrying you. This is the one genuinely hard technical requirement. Expect to be tested on it.

• AI experience in one of three shapes: working AI-natively with judgment; building AI infrastructure rather than calling an API; or AI enablement inside a company — internal tooling, harnesses, engineered context.

• Range across the stack. Front-end lean preferred. Hard specialists in any single layer are a no, infrastructure-only included.

• IC, not manager. There is nobody to manage and they've been burned by senior candidates who wanted reports.

Years of experience: roughly three to eight is the centre of gravity, but the client has been explicit that it's soft in both directions. If you're outside it and the rest fits, talk to us.

Helpful, not required

Health tech or pharma tech. Fintech as a partial substitute. Playwright or browser automation. Anything involving regulated data and models — that's the profile they get most excited about.

The parts people find hard

• Five days a week in the office. Not hybrid, not becoming hybrid. If you're going to want to renegotiate this at offer stage, tell us now — the client has been burned on exactly that and it's the fastest way to lose an otherwise good process.

• Not a nine-to-five. Their words. Also not 996, and they don't work weekends.

• No geographic pay premium. They'll pay a relocation stipend, but they won't raise the salary because you're moving from a more expensive city.

A note on non-traditional backgrounds

Job hopping, a gap, a layoff, no degree, self-taught, a company that folded — none of these take you out of the running with us, and we've confirmed with this client that none of them are filters on their side either. One of their favourite profiles so far was someone who joined a startup at 25 employees, watched it scale to 100 and shrink to 31, and stayed to the end. Tell us the story; we'll take it to them with the context attached.

US-based candidates only. We don't submit candidates who aren't already in the United States.