Most fractional CTOs hand you a roadmap and a hiring problem. This engagement embeds CTO-level judgment in your company, with the senior team that ships what gets decided. Monthly retainer plus scope, long-running by default. Closer to a co-founder than a vendor.
The default is both together, because that is the version that removes the most risk. The other exists because not every founder needs the same thing, and quoting you for a team you already have is not a service.
Judgment only
Wojciech works part of your week on architecture, hiring, vendor calls, and what not to build. You keep whichever engineers you already have. Right when there is a team that needs direction and no delivery gap to close.
The default
The judgment and the hands, so decisions get made and then shipped by the same people who made them. This is what most engagements are, and what we would quote unless you can name why one of the other two fits better.
We sell the judgment without the hands, or both together. We do not sell the hands on their own: capacity against a spec you hand us is the agency model, and nobody on our side would be positioned to tell you the spec is wrong. Both together costs more and removes that exposure, which is why it is the default rather than an upsell. Both are quoted after a scoping call. The audit is the only fixed price on the site: €4,900, fixed scope, report in 7 business days.
The default agency model assumes a known scope and a clean handoff. Domain-expert founders rarely have either: the scope keeps moving as you learn what your market actually wants, and there is no in-house engineering leadership to hand anything over to. So the engagement is built around those two facts rather than in spite of them.
The CTO
Architecture, hiring, what to build next, what not to build. Wojciech is in the room when those calls get made.
The team
Three more senior developers behind the CTO, working to the same standards. Real capacity to ship two workstreams in parallel.
The pricing
You're paying fees for ongoing technical judgment, and the cap table stays yours. Scope shifts as you learn. The model is built for that.
The horizon
We stay long enough to ship, then long enough to maintain what we shipped. Most engagements run multiple years. And when you're ready to hire in-house, we help you make those hires and hand over a codebase they'll thank you for. Reaching the point where you don't need us is a good outcome, not a failed one.
Satellite data platform
Two founders with deep space-domain expertise and, when we met, very little software between them. Mid-build and stuck on the architecture for a satellite data platform. The first real call went against the fast path: an existing open-source service would have worked for the modeling core, and we recommended skipping it and building a lightweight parser instead, because that was quicker to validate and easier to reshape once real engineers were using it. They made the call, we built it, and it’s been in production since 2026.
Started as advisory. Converted to a partnership engagement once the technical answer was clear. Founders kept ownership. We kept technical judgment.
Political finance, regulated industry
A non-technical political-finance operator brought us a decade of regulated-industry expertise and a problem nobody had solved well. We took it on as our own, and Wojciech is a co-founder. Bank sync shipped in the first sprint on whichever provider spiked fastest, but behind an abstraction built on the assumption that the first vendor pick would turn out to be wrong. Nine months later it was, and swapping the provider never touched the features sitting on top of it.
Architecture from scratch. Co-founder-grade equity stake. Long-running engagement, still active.
“Embedded technical partner” is the kind of phrase every agency writes, so here is the specific version. Four things are true of every engagement we run, and they are the things you would notice in the first month.
Decisions
Before anything gets built, the choices go into a written proposal: what each one costs, what it locks in, what it rules out later. That document is the thing you argue with, and it’s deliberately easier to change than code is. On Systellar, the access-control RFC sat in the repo five months before the code that implemented it.
Disagreement
On that same access-control decision, the founder read the proposal and argued that a whole layer of it wasn’t needed. He knew the use cases better than we did, he was right, and the design got simpler. If you never overrule us on something, one of us isn’t doing the job.
The room
Architecture, hiring, what to build next, and what to stop building. Not a written summary afterwards, and not a junior relaying a decision made somewhere else.
The record
Decision records live in the repo next to the code they explain. When you hire in-house, or bring in another team, or come back to a choice two years later, the reasoning is still there. This is the part that’s worth the most and gets skipped the most.
What it costs you: this is slower at the start than a team that takes a spec and begins typing. The first weeks of an engagement produce more writing than software. If what you want is someone to disappear with a Figma file and come back in six weeks, you will find this shape frustrating, and you would be right to.
How we work
It writes a good deal of our code. What it has never done is decide what to build, or notice that the thing being built was the wrong thing, and that gap is the entire reason this engagement exists. The discipline that makes the speed safe is a longer argument than this page should carry: how the type system catches invented APIs before they compile, what we put through review, and why we write the specification first. It’s all set out on the stack page.
Not sure whether you want us, a co-founder, an agency, or your first senior hire? The guide compares all of them, including the cases where we are the wrong answer.
Tell us what you’re building and what you know about your field that nobody else does. We read every message and usually reply within a day.