AI engineer,
product builder,
pragmatic by default.

I help founders, operators, and engineering teams turn messy ideas into shipped software, cleaner systems, and decisions that hold up in daylight.

No theatre. No corporate varnish. Just thoughtful execution, technical range, and a bias toward things that actually work.

What I do

Built for real-world work: product strategy, frontend craft, backend implementation, automation, AI workflows, deployment, debugging, and the ugly in-between bits nobody puts on the landing page.

How I help

Useful where software meets momentum.

Build new things. Product pages, apps, internal tools, prototypes, and production systems with taste.
Unstick old things. Legacy cleanup, bug hunts, architecture triage, and getting half-finished ideas back on rails.
Make AI actually useful. Workflows, assistants, automation, retrieval, evaluation, and sensible human-in-the-loop design.
Sharpen the thinking. Technical direction, product trade-offs, implementation plans, and honest second opinions.

Working style

Direct, opinionated, and allergic to fluff.

I prefer crisp goals, honest constraints, and momentum over rituals. If something is vague, I’ll try to make it concrete. If something is broken, I’ll trace it to the root rather than decorating the symptoms.

Stack

Wide enough to be useful.

Rails Ruby JavaScript HTML / CSS Turbo Stimulus SQLite GitHub GitHub Pages Figma AI tooling Automation CLI workflows AWS

Get in touch

If you need technical help that cuts through fog, say hello.

{/* Email contact removed at owner’s request — no contact method shown */}

Writing

Notes for operators, builders, and people trying to make AI useful without making it theatrical.

OpenClaw and Paperclip housekeeping Backup bloat, retention, verification, log rotation, and how to make a local ops setup boring again.
Good prompting, memory, context, and the graphs A plain-English explanation of prompting, context, memory, and how to think about inputs, outputs, cache reads, and cache writes.
A practical prompting cheat sheet The prompt formula, token-usage heuristics, and a practice model you can actually use without turning this into a cult.

New on site

Three new pieces, written like they belong in the wild rather than trapped in a notes app.

If you want the Medium-style version of the ideas we worked through — clean, readable, and fit for sharing — start there.

Read the latest Dev anthem

AI engineer, product builder, pragmatic by default.