I build infrastructure for AI workloads, developer tooling, and healthcare data systems from the mountains of southwest Virginia.
Most of my work lives at the intersection of Kubernetes, AI/ML operations, and protocol design. I like building things that make complex systems feel simple - whether that's scheduling GPU workloads across a cluster, wiring up MCP servers across editors, or transforming messy healthcare data into something useful.
fi-fhir - Healthcare integration platform that transforms legacy formats (HL7v2, CSV, EDI X12) into semantic FHIR events. Includes a mapping studio UI for building transformations without writing code.
edilint - A pre-send linter for EDI and healthcare flat files. Single static binary with CI-friendly exit codes: X12 envelope integrity, lookalike-character detection, declared-vs-actual count verification. Zero dependencies.
FlexInfer - Kubernetes-native GPU scheduling and AI model serving. Automatic hardware detection, intelligent workload orchestration, dynamic LoRA adapters, and multi-backend support (Ollama, vLLM, llama.cpp, and more). Built in Go.
MentatLab - Mission control for composable AI agents. Visual DAG canvas for designing workflows, Kubernetes-native execution, and a simple stdin/stdout protocol so agents can be written in any language.
Loom - An MCP (Model Context Protocol) ecosystem spanning a Go daemon, VS Code extension, and Zed plugin. Manages MCP server lifecycles, syncs configuration across Claude, Codex, Gemini, Kilocode, and other AI coding tools, and provides persistent agent memory across sessions.
I run a homelab K3s cluster on Harvester with Flux CD GitOps, Longhorn storage, and a full observability stack (Prometheus, Grafana, Loki). It's where I develop, test, and dog-food everything above.
My typical stack:
Go · Python · TypeScript · Rust
Kubernetes · Flux CD · Harvester · Longhorn
Prometheus · Grafana · Loki
MCP · FHIR · HL7v2
I contribute to OpenHands (WebSocket/HTTP proxy for K8s deployments) and maintain a collection of shared libraries across Go, Python, TypeScript, and Rust - from MCP SDKs and circuit breakers to semantic code indexing and browser automation.
I'm always interested in conversations about AI infrastructure, healthcare interop, or homelab builds.