The Workbench

A public workspace for things I’m building.

The Workbench is where I keep work I want to share: products like ReadyRoll, public GitHub projects like the Data Agent Suite, writing, and experiments. Some of it is commercial, some of it is for learning, and some of it is simply worth putting somewhere people can see it.

Product

ReadyRoll

A no-account scheduling app that helps groups find the best date to meet.

ReadyRoll started from a small coordination problem: groups want to meet, but finding a date often turns into a long thread of maybes and reminders.

Public GitHub Project

Data Agent Suite

A completed 10-agent suite exploring practical data workflows, deterministic evidence, bounded LLM use, and review-ready artifacts.

The suite follows a deliberate path: start with local deterministic checks, add bounded LLM support only where it helps, then build toward review workflows with traceable outputs. Each agent is a practical data tool first, not a chatbot demo.

Progression

Deterministic checks Bounded LLM support Review workflows Orchestration seams 10 completed agents

Working Principles

  • Start with the use case

    Understand the purpose before choosing the method, tool, or architecture.

  • Question the obvious answer

    Best practice is only useful when it fits the problem in front of you.

  • Build through experiments

    Prototype, test ideas, and let the work teach you what should come next.

  • Keep only what matters

    Aim for the point where removing anything else would make the thing worse.

  • Clear traces, reports, and documentation

    Make the work visible so others can follow and verify it.