The version before this one pitched client work and nothing else. It is still true — I take that work — but it stopped being the whole story a few weeks after I wrote it.

Here is the rest of it. I am a solo developer in Vietnam. I build self-serve products for merchants, agencies and bookkeepers in the US and UK, and I take full-stack React, Node and TypeScript client work alongside them. Support is async — email, not calls. If a product needs a call to be usable, it is not finished.

In September 2026 I stopped guessing. On the 4th I locked eight products — two Shopify apps, a WordPress plugin, two iOS apps, two web SaaS and a Mac app — to be built in parallel rather than in a queue. The next day I tried to knock them out: 96 challenger ideas, scored against the same weighted rubric, with 30% of the weight on one question — can this reach its first hundred paying customers when nobody knows who I am? None of the new ideas reached the bar, so the eight stayed.

When I locked them, one had real code behind it and the rest had locked specifications and scaffolded repositories; by the middle of the month all eight had real code, and none had launched. Each one carries a written condition under which I stop working on it.

What has not changed is the bet: be useful in public long enough that the right people find me already convinced. What changed is that I now have evidence to publish instead of intentions.

Full-Stack Developer. React, Node, TypeScript. AI and LLM integration. Remote for US teams.

Three-plus years building full stack, and the part worth naming is the narrow one: taking AI from “we should use it” to something running in production, with the error handling, the cost ceiling and the fallback path that a demo never needs. I use Claude and OpenAI tooling every day — as the systems I build, not as autocomplete.

I own features end to end, from database schema to UI. I work written-first and async-friendly: clear specs, small pull requests, honest status updates. If you need someone on a call to know what I did this week, I have not written it down well enough.

I do not work with WordPress, PHP or Angular.

This version is the pitch as it stood in August 2026, before the product portfolio existed. It is still accurate for client work — it just stopped being the whole story a few weeks later.