I spent more than 15 years in demand and supply planning at Sears, Belk, Apple and Anker, then started building the AI tools I wished I'd had. Now I run Bartlett Labs, where I build AI agents and automations for operators and small businesses.
Most of what I build runs in production on my own servers. A few highlights:
| Project | What it is | Try it |
|---|---|---|
| RepBot | AI receptionist that answers missed calls and books jobs for local service businesses. A product I sell, so this repo is a case study | Site |
| Supply Radar | Supply chain risk dashboard you can question in plain English, with what-if scenarios drawn on the map | Live demo |
| Weigh | Before/after eval harness for LLM pipelines: binary checks, cost per case, confidence intervals, a CI gate | Live |
| PayGuard | Reads a pay stub with an LLM, then checks it against federal and state wage rules in plain code | Live demo |
| agent-bus | MCP and REST message bus that lets my Claude Code agents and Grok bots message each other directly |
How I build: TypeScript first, small services I can deploy and watch myself, and evals before I trust a model with anything that touches a customer.