Insights
I write about enterprise AI, ERP architecture, and the messy reality of making large-scale technology work. No fluff — just what I've learned building real systems.
Stripe ships 1,300 PRs every week using parallel AI coding agents. Here's the full architecture — Claude Code CLI, git worktrees, OpenClaw orchestration — and how I replicated it on a single server in an afternoon.
Read Article →Most AI ROI claims are aspirational. This one is documented. Full breakdown: the investment, the 5 measurable outcomes, the methodology behind the number — and what IT leaders consistently get wrong about the investment decision.
Read Article →Not a tutorial. A retrospective on deploying 58+ specialized AI agents using the BMAD methodology, the context optimization system that saved 70-85% on tokens, and why most enterprise AI initiatives miss the mark.
Read Article →Building a RAG system is easy. Building one that works at 86% accuracy with 22ms latency in a production enterprise environment — that's a different challenge entirely. Here's what I learned deploying Qdrant + Neo4j at scale.
Read Article →Board-level technology investment decisions aren't won with demos — they're won with data, risk models, and honest ROI projections. How 65 days of production log analysis changed everything.
Read Article →