Who we build with
Teams we build with
Golux is a founder-led practice, so we are specific about the work rather than the logos. Below is the kind of team we do our best work for, the products we have built and run ourselves, and the stack behind them. Client references are real people who will take a call. We introduce them when a conversation gets serious, with their permission, rather than printing quotes on a page.
The work we are good at
Four situations we keep being called for
An AI-built MVP with real users
Shipped on Lovable, Bolt, v0 or Cursor, now carrying customers and risk. We audit it, fix what is dangerous, and take it to production without stopping the product.
A backend nobody wants to touch
It works, and every change is a gamble. We draw the boundaries, write the tests that make change safe, and hand it back to a team that can ship into it weekly.
An AI feature that has to be real
Past the demo: grounded in your own data, measured against real cases before launch, with a cost ceiling and a human approval step where the risk sits.
A product team that is one or two people short
Senior engineering or design judgment inside your process, for the hours you need, not headcount you have to hire for.
What we can show you today
The strongest reference we have is the software we run ourselves
Every product in the portfolio is live and operated by us: server monitoring, email infrastructure, AI matching, AI marketing, project forecasting, and wellness and education. Open the project pages to see what each one does, then tell us which experience is closest to your problem.
Technologies we use
Mainstream on purpose. Everything we build is something your next engineer can be hired for, and nothing here requires us to stay.
- Ruby on Rails
- TypeScript
- Python
- PostgreSQL
- REST APIs
- Docker
- React
- TanStack Start
- Next.js
- Tailwind
- Vite
- React Native
- Progressive web apps
- Figma
- Design systems
- Prototyping
- GA4
- Search Console
- Google Tag Manager
- Technical SEO
- OpenAI API
- Anthropic API
- Retrieval (RAG)
- Evaluation sets
- Agents with human review
- pgvector
- SQL
- ETL pipelines
About references
We do not publish client quotes. The ones worth having are conversations, and the people who would give them are busy, so we ask them when a deal is close, for a specific person, rather than collecting testimonials for a web page. If you want to speak to someone we have worked with before you commit, say so on the call and we will arrange it.
Next step
Start with something small and see for yourself
The audit is free and takes a minute. Naming one problem costs a short call and nothing else. Neither requires you to believe anything we wrote on this page.
