We registered TEKKNA on 1 January 2026. The choice of date was less about symbolism than about logistics — German notaries don’t take walk-ins on 31 December — but the idea behind the shop had been waiting much longer.
This is the short version of why a new engineering studio in Hamburg made sense in 2026, what we set out to build, and what the first day actually looked like.
Why a new shop, why now
By late 2025 a few things had stopped being controversial:
- AI had become load-bearing in how software gets built. Not as a replacement for engineers, but as a power tool that materially changed the cost of writing, refactoring, and reviewing code. The teams treating it as such were pulling away from the ones still arguing about whether it counted.
- EU digital sovereignty stopped being theoretical. Schrems II hadn’t been repealed. The CLOUD Act hadn’t been rescinded. Procurement teams in Europe had started asking the hard questions in writing, and the “we’ll figure it out later” answer had stopped working.
- Small senior teams ship faster than large junior ones. They always have. The change in 2025 was that the cost of running “small and senior” had dropped — better tools, better infrastructure, AI as an extra pair of hands — to the point where the model became more attractive than the alternative.
We wanted a shop that took all three of those seriously, in one place, on European infrastructure, with no layer of project managers between the customer and the people writing the code.
The model
The shape of TEKKNA is deliberately opinionated:
- Senior engineers only. The person who pitches you is the person who builds. Nobody is handing your project off to a junior on a different floor.
- B2B software, four lanes: greenfield product engineering, cloud architecture, EU sovereignty migrations, and production AI agents. Not a long menu — a short one we can actually be good at.
- EU-resident infrastructure as a default, not a checkbox. If a workload can run in Europe, it does. If a customer needs it to, we make sure it can.
- One small in-house product. Atheena — AI-supercharged customer support for B2B SaaS — so we eat what we cook and feel our own latency, our own incidents, our own bills.
What 1 January actually looked like
Nothing dramatic. A notary appointment we’d booked two months earlier. A bank that wouldn’t open the business account until 5 January. A first commit to a private repo at 22:17 CET — TypeScript, two files, no tests yet — the kind of unceremonious first move that doesn’t look like a beginning until much later.
We were operational on paper in January. We took the first paid engagement in February. The website you’re reading is its own small ship — written, deployed, and updated by the same people doing the engagement work.
Four months in
The first stretch is intentionally narrow:
- Two greenfield B2B SaaS builds in flight.
- One cloud architecture engagement on Cloudflare.
- A Go Europe migration off a US hyperscaler.
- Internal Atheena work, with live customers on EU-resident infrastructure.
It’s a short list and meant to stay that way. We’d rather six well-shipped engagements in a year than thirty half-shipped ones.
What “less friction, more impact” means in practice
The tagline isn’t a slogan, it’s an operating instruction:
- Less friction means short feedback loops — small teams, direct communication, automated deploys, no decision waiting on someone in a different time zone.
- More impact means picking work where the engineering meaningfully changes the outcome. Not maintenance contracts. Not staffing gigs. Things that ship and matter.
What’s next
The shape of 2026 from here:
- More Go Europe work. Procurement is moving faster than most people realise. The conversations we’re having now are about replatforming, not benchmarking.
- More agentic work. Not demos. Production systems with retrieval, evals, and SLOs — the boring loop around the model that’s the actual product.
- More writing here. This archive will get less sparse. Most of it will stay short.
If you’re an engineering leader, a founder, or someone in procurement reading this and the model resonates — we’d rather have one good conversation than ten cold pitches. stefan@tekkna.dev.