Embedded, backend, infrastructure — the whole path, one engineer
LEMB Tech is my engineering practice. I started in DevOps — building and operating the infrastructure other people's software depends on — and over the years began building that software too. Now I take work the whole way: the code, the service around it, and the infrastructure that keeps it running.
Embedded, backend and infrastructure carry equal weight in that. A device that can't reach a dependable service is a prototype. A service nobody can deploy or keep up is a repository. Being able to cross those boundaries without a hand-off — and without waiting for someone else to have time — is the whole point.
Everything under Projects is mine end to end: designed, written, packaged and deployed by me, running on servers I administer. That page is the argument; this one is just the summary.
Where I started, and still the foundation. Linux, Docker and Compose, reverse proxies and TLS, CI/CD, backups, monitoring, secret management. Not a diagram of a platform — one I keep running, and get paged by when it doesn't.
Rust, Python, C/C++, and the web front ends that sit on top. REST and WebSocket services, data pipelines, browser tools that do real work client-side. Written to be deployed, because I'm the one deploying it.
PCB design, schematic capture, firmware. Boards built around the constraints of the job rather than adapted from a reference design — and firmware written by the person who drew the schematic.
The three above are worth more together than apart. A sensor, the service that collects from it, and the machine that service runs on are one system — and most of the expensive problems live exactly on the seams between them.
The quickest thing I can show you is the page you already have open. It is served by the same setup as everything else on this domain, and it was built the way I think a site of this kind should be built.
One engineer. I came up through DevOps and infrastructure, which is still where my instincts are — I think about how a thing will be deployed, watched and recovered before I think about how it will be written. Development came later and stuck, so I now build the things I used to only run.
That means one point of contact, no translation layer between people who don't share a vocabulary, and decisions made by whoever has to implement them. The CV has the detailed background; the projects show what it produces.
