Hands-on technical executive with 30+ years of experience building, scaling, and rescuing complex healthcare and enterprise systems. I operate across all levels — from writing production-grade code to leading engineering organizations — and have spent the last few years making LLM-based development reliable enough to ship.
I started in healthcare IT in the early 2000s — SVP of Operations and then Chief Architect at MediNotes, then VP of Operations at Eclipsys — working on electronic health records and practice-solutions systems used by physicians and nurses every shift. It was the era that taught me the only thing that really matters about software in healthcare: it has to work, every time, or someone gets hurt.
The 2010s were all leadership. Five years as CTO at BettrLife, a mobile-health startup, where I built the team and the product from the early days through scale. Then six more years running Hills Consulting LLC — fractional-CTO work for a string of healthcare-tech companies (Certintell, Redivus Health, Play-It Health, MyDiabetesHome, ConnectRN) that needed experienced leadership without a full-time hire. Then ConnectRN itself, where I came in as Principal Engineer and left as Technical Director / Lead Architect — the primary escalation for production incidents across a high-traffic nurse-staffing platform.
Most recently: CTO at Alopex, back in healthcare IT — rebuilding a fragmented
platform into a unified Go / React / AWS architecture. On the side, quietly
building the AI coding tools I wished existed. Those projects
(speccritic, plancritic, prism, realitycheck,
clarion, verifier, atlas) became a connected
AI-Native Development Toolchain — and are now most of what I work on.
Now: independent again. Hills Consulting, full time. Building open-source tooling. Writing on Medium. Available for the kind of work that values both speed and rigor — and increasingly skeptical of anything that promises one without the other.
The most expensive part of software is the part that's wrong. That's true with humans; it's triply true with LLM agents. Most of my open-source work starts from the same premise: turn requirements into a contract, check the contract, then write the code.
Everything I ship is small, opinionated, and designed to be composed. One command does one job and returns a deterministic exit code. If you can automate it, a CLI beats a UI. If you can explain it in a README, you probably don't need a framework.
Twenty years in healthcare IT is a crash course in the difference between "the tests pass" and "the system actually works." Software that can't afford to be wrong teaches you a different kind of discipline. I try to bring that discipline to everything now, not just medical systems.
I publish essays because I don't trust ideas I can't explain to a stranger. The writing is the thinking — what's clear in prose tends to be clear in code, and what's muddy in prose almost always produces muddy systems.
Hands-on CTO work for companies past product-market fit but facing the architectural, team, and delivery bends that come next. Recurring 1–3 day/week engagements. Distributed teams, roadmap ownership, hiring, and the unglamorous work of turning ad-hoc engineering into a shippable practice.
Deep reviews of existing systems, cloud infrastructure, and team practices — grounded in two decades building HIPAA-compliant healthcare platforms (HL7, FHIR, EHR integrations). Particularly valuable before a migration, a platform rebuild, or a diligence round where the technical story has to hold up.
Helping teams move from prototype LLM features to systems you'd actually ship: spec-first workflows, deterministic review, context budgets, evaluation, observability, cost. The working material behind my open-source toolchain and most of the writing.
Board advisory, technical diligence for investors, and focused workshops on AI-native development for engineering teams. For when the question is narrow, the stakes are real, and the answer benefits from someone who's been shipping software for thirty years and writing about the transition.
I'm currently booking engagements for 2026. If any of the above sounds like a fit, the fastest way to start is a short note describing what you're working on.