Technical product leader. I spent eleven years at Global Payments building developer experience infrastructure at enterprise scale: PCI, tokenization, 3DS, ACH, the parts that matter when money moves. Now I'm building LeadSurface.
Most of my personal work lives at the intersection of AI tooling, web standards, and developer experience. Louisville, KY.
LeadSurface turns forum noise into qualified pipeline. It monitors Reddit, HN, Discourse, and StackExchange for competitor switching signals and routes them to sales before the prospect is gone.
herdbot runs persistent AI bots on herdr. Each bot gets a workspace, subagents run in panes, and tasks route across your Claude, Codex, Cursor, and Pi subscriptions. Written in Rust.
Loop & Gate is the system I build LeadSurface with. The foundation gives an AI agent files-first memory in an Obsidian vault, and three kits add human gates for building, going to market, and following through.
These exist because I needed them and nothing adequate existed:
| claude-code-config | My full Claude Code environment: hooks, skills, plugins, permissions. Public because other people asked. |
| deferred-skill-catalog | Keeps Pi and Claude Code skill catalogs out of the prompt and searches them on demand. |
| gliner-browser | A local browser CLI for AI agents, with persistent Playwright sessions and GLiNER2.5-powered element targeting. |
| bb-tui | A terminal UI for bb. Browse threads, follow streaming agent output, and reply without leaving the terminal. |
| skills-writing | Writing skills for Claude Code that strip AI tells from prose and enforce the Microsoft Writing Style Guide. |
| skills-design | 23 brand-agnostic design templates for Claude Code: social, blog, YouTube, podcast. |
| skills-vault-knowledge | 20 Obsidian knowledge-work skills: connect ideas, trace evolution, surface contradictions, ghost-write in your voice. |
| hermes-dispatch | Local-first agent dispatch for the Hermes AI harness, reachable from your phone over Tailscale. |
| mac-mini-llm-roster | Benchmark data for 20+ local LLMs on 32 GB Apple Silicon: tokens/sec, wall-clock, quality. |
- Token Routing Is Least-Cost Routing —
- Your Agents Made You Faster. Your Customers Didn't Notice. —
- I Mined My Session History for the Prompts I Actually Reuse —
- The AI Implementation Playbook, Annotated —
- Value Creation Just Became the Distribution Strategy —