Internal Tools & Projects

A collection of building principles, internal tool philosophies, and pragmatic software experiments.

// What I Focus On

src/focus/support-ops.md

Support Operations & Internal Tools

Eliminating repetitive toil in operational environments. Workflow orchestrations, triage tooling, knowledge lookup enhancements, and pragmatic automations for support teams.

Solving real user friction at the operational coalface, not inventing abstract architecture.
Internal dashboards, command-line helpers, and automations that save hundreds of human hours.
Designing systems that respect human attention and fail gracefully.

src/focus/applied-ai.md

Applied AI & Intent-Driven Building

Leveraging modern LLMs and agentic toolchains to build software without a traditional computer science degree. The entry barrier has collapsed; intent and taste are what remain.

Wrapping stochastic models in deterministic schemas, typed validations, and clear invariant tests.
Using Model Context Protocol (MCP) and agent loops to turn natural language into reliable operational glue.
Prototyping fast, testing in the open, and ruthlessly deleting unnecessary code.

src/focus/nerdery.md

Life, Hobbies & Nerdery

When I am away from the terminal, I am raising my family, playing Magic: The Gathering, and passionately yelling at the TV during Pittsburgh Steelers games.

Fatherhood: The ultimate non-deterministic real-time system.
Magic: The Gathering: Appreciating Richard Garfield's magnum opus as both a competitive card game and an elegant priority queue.
Pittsburgh Steelers: Unapologetic Terrible Towel waving, defensive slugfests, and heart rate spikes every Sunday.

// Principles for Building Internal Tools

Building internal software for support operations is fundamentally different from building public consumer apps. Here are the rules I live by:

1. Solve the Actual Friction, Not the Theoretical One

It’s easy to dream up complex dashboards that nobody visits after week two. The most transformative internal tools are often invisible: a CLI script that formats messy merchant payloads, a Slack bot that pulls relevant telemetry during escalations, or an automated webhook that catches broken states before a human has to diagnose them manually.

2. Keep the Surface Area Boring and Fast

Internal users are trying to get their jobs done quickly. They don’t want 3-second splash screen animations or experimental design paradigms. They want keyboard shortcuts, instant search, copy-to-clipboard buttons that always work, and zero layout shifts.

3. Fail Gracefully & Loudly

When an internal automation breaks, it shouldn't fail silently or pretend everything succeeded. It should print a crisp, diagnostic error message, preserve the user's input so nothing is lost, and provide an obvious fallback path.

4. The Best Code is Deleted Code

If an operational process changes and a tool is no longer serving a purpose, deprecate it aggressively. Tools that outlive their utility become ghost ships that confuse new team members.

// Connect & Chat

Building something similar in ops or internal tooling?

Drop a line or read more in the blog notes.