Open-source workflow
The Vault
A persistent, LLM-maintained wiki that connects product decisions, engineering implementations, raw evidence, and the context agents need to make better decisions.
Senior Software Engineer · Tri-Cities, Tennessee
I build production AI systems and the knowledge infrastructure that helps humans and agents understand them.
I'm William “Billy” Bratz. My work sits where LLM agents, MCP infrastructure, distributed systems, and engineering judgment meet. I build systems that survive production, then turn what they teach us into context that compounds.
Selected work
Systems, tools, and ideas made concrete.
Production systems
AI agents and MCP
Building and operating the tool, context, evaluation, and reliability layers behind customer-facing conversational AI.
See the engineering recordEngineering practice
Theory-building series
Three essays about why software understanding lives beyond code and documents. I also write about what changes when the map can learn.
Start with part oneSelected writing
Notes from building, operating, and thinking.
When the Map Starts Learning: Building a Shared Memory for Humans and AI
What happens when documentation stops being a destination and becomes living context for product, engineering, and AI.
The Map Is Not the Theory: Why Docs and Diagrams Will Mislead You
Written artifacts are just approximations; the real system lives in the team's head.
Programming as Theory Building: Why "Just Read the Docs" (RTFM) Never Works
Why understanding code requires more than documentation alone.
About
Production is where the real system introduces itself.
I've spent 11 years building distributed systems across identity, underwriting, analytics, and conversational AI. I care about evidence, explicit decisions, operable systems, and the hard-won theory teams develop while making software work in the world.
Today I'm focused on production agent systems and on preserving the connection between why a product decision was made and how it was implemented. That is the idea behind The Vault and much of what I write about here.
Let's compare notes
Interested in production AI, MCP, or engineering knowledge systems?
I enjoy comparing notes with people working through the same hard problems.