A natural match if you have experience building agentic AI systems for high-stakes domains.
Derived from job-description analysis by Serendipath's career intelligence engine.
Original posting from Foundation Health via Ashby
About Foundation Health
Foundation Health is an AI operating system for pharmacy - built by a team who lived the problem firsthand, not one that came to healthcare from the outside. Every day, prior authorizations delay prescriptions, pharmacists lose hours to paperwork instead of patients, and prescriptions slip through the cracks between health systems. We believe thoughtfully applied AI can finally change that - deep, integrated infrastructure that streamlines how care is coordinated, purpose-built for the complexity of specialty pharmacy.
We're building toward a world where pharmacy teams spend their time on patients, not paperwork - where AI handles the routine so clinicians can focus on what only humans can do. Getting there takes the right people. We're looking for curious, ambitious folks who want to help solve healthcare's hardest problems - for our patients, our partners, and each other.
The Role
This isn’t a pitch deck. Our platform is live in production inside major US health systems. Our AI handles real patient communication over voice and SMS and plugs directly into Epic workflows. If it gets something wrong, there’s no funny screenshot: a real person doesn’t get their medication. That’s the standard we build to.
You’ll own architecture across several product areas and design systems that can grow 10–100x without a rewrite. When the hardest problem in the company comes up, you’re the one who opens the editor and builds the first version.
You’ll also make the rest of engineering better. That means mentoring, leading projects across teams, and changing how we build, including how we use AI to ship faster without shipping worse. If you’ve already rolled out agentic dev practices at a team or company level and can show what happened to output and quality, we’d love to hear about it.
You’ll work directly with our VP of Engineering (ex-DeepMind) and CEO. Nobody here manages from a distance, at any level, and that includes the Principal.
The problems you’d own
- Agentic AI that’s safe to put in front of patients. This covers voice and SMS agents, plus the evals, guardrails, red teaming and pharmacist-oversight design that make them trustworthy. If you’ve worked on AI safety or LLM security in production, this is where that experience pays off.
- Deep EHR integration. Epic queues, prior authorization automation and 340B compliance, where correctness really matters and the specs are older than some of our engineers.
- Our target-state architecture. You’ll take the platform from a handful of health systems to hundreds and build the first mile of that migration yourself.
- How we do AI engineering. That includes model selection, context management, cost and latency tradeoffs, and evaluation methodology.
Who tends to do well here
- You’ve built serious systems with your own hands recently, and you can go deep on the details of anything you’ve shipped. Your title doesn’t matter much. Plenty of great Principals have spent time as directors or heads of something without ever stopping building.
- You’ve worked at Staff+ level somewhere with a high bar, and you can point to architectural decisions a company still runs on.
- You’ve built with LLMs in production, well beyond SDK calls: agents, tool use, evals, guardrails, and the unglamorous work of making them reliable.
- You keep things as simple as the problem allows. You’d sooner run a well-structured monolith on Postgres than ten microservices nobody can debug, and you change the architecture when real constraints show up, not before.
- You’ve worked somewhere mistakes are expensive. That could be healthcare, but fintech, fraud, trust and safety, or anywhere compliance shapes design counts too. We don’t need you to know healthcare yet. We need you to care about getting it right.
- You find the highest-leverage problem in the company and solve it without waiting to be pointed at it.
Where we are
We’re about 70 people, past product-market fit, raised a $20M Series A led by Define Ventures, and our CEO is a 2x founder.
The practical stuff
We’re remote-first across the US, with an engineering hub in downtown Orlando that you’d visit from time to time. We’d ideally like you on Eastern or Central time, since a good chunk of our engineering team is in the UK and the overlap really helps. And it’s early enough that the equity is founding-role equity, not a rounding error.