Mateusz Juszczyk / Full-stack developer
I build reliable digital products across AI, cloud systems, and the modern web.
Explore workThoughtful architecture, direct interfaces, and resilient delivery shape every build.
A Nuxt 4 reservation platform backed by Cloudflare Workers, Hono, D1, and Drizzle ORM. Stripe Checkout handles payments, while database constraints protect apartment availability from double-booking.
Industrial design / keyboards
Before interfaces, there were objects: tolerances, materials, mounting systems, and the feel of a finished mechanism. These open-source keyboard projects turn CAD models into physical products you can build, hold, and modify.
Flagship / TKL / Open source
A triple-weighted TKL with a faceted Toblerone-style rear weight and three distinct mounting options. Clover is the most developed expression of the series: configurable, substantial, and designed for assembly.
First prototype / 60%
The first keyboard prototype in the series. A compact Bakeneko60-compatible case built around an O-ring mount and a two-piece weight that contrasts oxidised and raw brass.
Manufactured / 60%
A manufactured 60% case with a shark-themed weight, combining a Bakeneko-compatible O-ring layout with a deliberately simple three-metal construction.
Open to junior AI engineering and full-stack product roles
I study IT and Econometrics, where statistical rigour meets systems thinking. Beyond the lecture hall I build production-grade applications across the full stack, from reservation platforms and commerce checkouts to AI-powered communities and offline-first PWAs.
My hands-on experience with large language models, agentic architectures, and prompt engineering gives me a practical edge in applied AI. I follow emerging models and research with genuine enthusiasm, always testing new capabilities against real product constraints.
I believe the best software is purposeful, performant, and invisible when it works. Every project I ship is an exercise in that belief.
Currently exploring
I keep a short list of topics I am actively pressure-testing against real product constraints.
Designing reliable multi-step AI systems that stay inspectable and useful in production.
Pushing more product logic to the edge without losing clarity, observability, or cost control.
Building interfaces that stay calm and trustworthy when the network disappears.
I move from product decisions to production code without losing sight of the experience.
A good project starts with clarity
Tell me what you are trying to make.