hi, I'm Clint.
I build products.
I write down what I learn building them: healthcare integration, applied LLMs, and leading the teams behind it. Twelve years of it, mostly in Nashville, usually with espresso.

the one I built for the course →
A caddie that speaks first.
MetaCaddie reads the hole, the wind and your own game, then tells you the club and the target out loud, in the voice of the caddie you pick. On iPhone, Apple Watch and Ray-Ban Meta.
iPhone first. Android is coming.

fresh off the desk →
8 min readHealthcare integration
AI Agent Audit Log in FHIR: AuditEvent and Provenance That Name the Agent
Through a SMART backend client, the EHR logs a service account, not your AI agent. Build per-run AuditEvent and Provenance that name the agent, model and run.
Read it →

Latest writing
All 98 articles →- HL7, FHIR, MirthLLM LOINC Mapping: Make the Model Pick, Not Remember
- HL7, FHIR, MirthStop Polling the EHR: FHIR Subscriptions and Event-Driven AI Agents
- Snowflake, graphsAI Business Process Mapping Should Output a Graph, Not a Diagram
- LLMs, agents, evalsAI Forward Deployed Engineer Agent: Its Riskiest Output Is a Default
- HL7, FHIR, MirthIs Text Messaging HIPAA Compliant? Only With a Gate Before send()
Where to start
pick the one that sounds like you, read them in order
You work in healthcare tech
You build with LLMs
Series
HL7 & Mirth from zero
Shipping LLMs in production
React state, then and now