Founding Engineer, AI RCM Healthcare Platform (TypeScript, Python)

About Adentris

Adentris (Y Combinator) builds AI-powered compliance and revenue integrity infrastructure for behavioral health providers. Our systems read real patient records, find clinical and regulatory risk, and put it in front of the people who can act on it. We run on production PHI for paying customers, so every design decision reaches actual patients.

We're also architected differently from most healthcare AI companies: all model inference runs inside our HIPAA-compliant Azure environment. No external LLM APIs ever touch patient data. That data-sovereign architecture is a core reason enterprise customers choose us, and you'll own it.

The founders

You'd be joining three founders, not just a company. Dmitry Karpov (CEO) is a second-time YC founder with a track record in B2B enterprise products and go-to-market. Sergey Yudovskiy (CPO) is also a second-time YC founder: he previously ran ElectroNeek (YC W20) as CEO, scaling it to ~$5M ARR across 30+ countries. He also comes from a family of physicians, which is why claims denied over documentation errors are personal, not abstract. Alex Odin (CTO) is a third-time founder who led AI at ManyChat ($140M+ ARR) and built consumer products used by 25M+ people; he owns the multi-agent architecture behind every Adentris module.

You'll work with all three of us daily, and your line to a decision is one Slack message long.

The role

You'll own the technical direction of the platform end to end. This is not a senior IC role with a bigger title. You decide how clinical data is modeled, how our AI pipelines are evaluated and trusted, and what the platform looks like in two years. Then you build it.

You'll sit in customer calls with Directors of Compliance, leave with a problem nobody has written down yet, and ship it within the quarter. When there's a hard call on architecture, or on whether a model output is safe to show a clinician, you're in the room making it.

This is right for you if you've been the technical center of gravity on something real, and you'd rather own an ambiguous problem than a well-specified ticket. It's wrong for you if you want a defined scope or a team that already exists. You're the person who creates them.

Your first 90 days

• First 30 days: ship to production. Take over one of our five live modules (Documentation QA is the busiest), get its evaluation harness under your control, and ship an improvement a customer notices.
• By day 60: own the EHR data layer. We integrate with the behavioral-health-native EMRs (Kipu, Alleva, BestNotes and others), and the ingestion and normalization architecture across them is yours to set.
• By day 90: you've made at least one architecture decision the company will live with for years, defined how we evaluate model output before a clinician sees it, and set the bar the next engineering hires will be measured against.

What you'll own

• Architecture across the full stack: data model, ingestion, AI pipelines, application layer, infrastructure
• The clinical data layer: ingesting and normalizing messy EHR data: FHIR, HL7, and the many formats that pretend to be them
• AI systems clinicians actually trust: evaluation, ground truth, error analysis. Calling a model is easy. Proving the output is right is the job
• Product judgment from data: go into real client data, find quality gaps nobody has flagged, turn them into shipped features
• The engineering bar: as the team grows, you define how we hire, review, test, and ship

What we're looking for

• 5+ years shipping production software, with real ownership of systems that outlived your involvement
• Deep TypeScript/Node and Python; advanced React and Next.js
• Data engineering at scale: schema design, query optimization, pipelines over large volumes of semi-structured records
• Production LLM systems, not demos. You've built the evaluation harness that kept them honest
• Communication that carries weight: equally clear with a clinician, a founder, and an engineer

Strong signals

• Founding or early engineer at a startup that reached real scale
• Regulated data experience (healthcare, finance, legal); you've been through a SOC 2 or HIPAA audit rather than only read about them
• Hands-on with EHR data and clinical formats; embeddings and retrieval in production
• Self-hosted or private model inference experience (Azure ML, vLLM, or similar)

Details

• Full-time, remote (US), with team meetups in San Francisco
• Competitive salary plus founding-level equity
• Direct line to the founders; short path from idea to production

Stack: TypeScript, Node.js, React, Next.js, Python, SQL/NoSQL, Azure-hosted LLM inference, Docker, Kubernetes

Interview Process

Our process is designed to be fast and mutual: you're evaluating us as much as we're evaluating you.

• Intro call with a founder (30 min). Your background, what you're looking for, how we work.
• Technical deep-dive (60-90 min). Walk us through a system you built and owned: the architecture decisions, the trade-offs, what you'd do differently now.
• Working session (2-3 hours). A realistic problem from our domain: reasoning over messy clinical data with an LLM pipeline, including how you'd evaluate whether the output is trustworthy.
• Founder round. Product judgment, ways of working, mutual expectations.
• References and offer. We typically move from first call to offer in under two weeks.