I spent two decades in enterprise operations -- the kind of work where you learn what breaks at scale, what keeps systems alive at 3 AM, and why the gap between what ops needs and what gets built is where projects go to die.
That operational foundation is why I build differently now. In 2026 I founded Stackbilt LLC in Texas to build AI infrastructure that actually works in production. The platform is live — Stackbilder and img-forge anchor a growing toolset (worker observability, CTO/CISO advisory, trust verification) on a unified auth and billing layer, with an internal MCP gateway for third-party client access. Real payments via Stripe; free tier is zero-cost -- 3 scaffolds and 5 images a month, no credit card.
The first thing I built was AEGIS -- a persistent autonomous agent that runs on Cloudflare Workers, routes across multiple AI models, maintains hybrid vector memory, and executes a full SDLC pipeline from GitHub issue to merged PR. It's not a chatbot. It's a cognitive agent that ships code, monitors infrastructure, and improves itself daily.
I distilled the architecture lessons from AEGIS and the wider Stackbilt platform into The Edge That Thinks, a public-review field guide to designing reliable generative AI systems.
Everything I build is edge-native by default, cost-aware by design, and shipped in small increments. Build the infrastructure, then build on the infrastructure.