I've spent 25 years teaching machines and teams how to work.
Director-level engineering leader with thoughts on technical leadership, systems thinking, and applying practical systems to new technologies for consistent results.
Before I managed engineers, I designed taxonomies and content models for hospitals and universities: the same discipline now called context engineering. I write about where that overlap actually holds up, and where it doesn't.
20+ years leading engineering orgs from 3 to 45+ people, across startups and global operations. Currently deep in RAG architecture, GraphRAG, and agent tooling: partly for the next role, partly because it's the most interesting problem in the field right now.
“Chunking is content modeling with a new name. GraphRAG is taxonomy that finally got a runtime.”
The Librarian Problem
From 2000 to 2007 I designed taxonomies, metadata strategies, and content models for hospitals, universities, and professional associations, back when "information architecture" was the term for it. Most engineers building retrieval systems today are learning those same lessons from scratch: how to chunk without losing meaning, how to build a hierarchy that a machine, so users can actually navigate under pressure.
I'm writing the full essay now, alongside a working RAG system built to test the thesis rather than just argue for it.
Recent Writing
Notes on engineering leadership, published on LinkedIn while the archive here gets built out.
Leading with Strengths
Read article →Top 5 Attributes of a Manager Employees Love
Read article →Ideas on Keeping Remote Employees Engaged
Read article →Currently Building
Two small products, built mostly by directing AI tools rather than hand-writing every line: a useful test of where that workflow holds up.