A curated set of systems, experiments, and internal tools I built to solve real delivery and operations problems.
Work in progress lives on GitLab and GitHub before it shows up here.
Follow on GitLab →Format-agnostic pipeline that parses legacy healthcare formats (HL7v2, X12 EDI, CSV) into semantic events and routes them through configurable workflows, with staged parsing via Source Profiles, FHIR R4 output, and CI coverage gates. The browser playground runs the real parsing and FHIR mapping kernel compiled to WebAssembly on synthetic samples; nothing you paste leaves the page. The Mapping Studio and remaining pre-1.0 work are described in the case study.
Healthcare parsing, transformation, and workflow delivery with explicit validation contracts
Pre-send linter for EDI and healthcare flat files.
AI-powered news pipeline.
Repository banner generation toolkit.
Intelligent diff analysis and patching.
Private runtime and inference control plane for Kubernetes.
Composable AI agent framework.
pip install fractal-agents
GitLab webhook event processor.
Kubernetes cost prediction and optimization.
Composable LangGraph agents for AI workflows.