Architecture, process, product, cost. And AI where it earns its place. Twelve years from engineer to CEO across fintech and trading. Point me at what's stuck; I find it, get it moving, and leave it running with the team you've already got.
Output that won't come. AI that won't land. Costs no one can explain. The instinct is to blame the nearest thing: a tool, a hire, a model. In my experience the real cause was set much earlier, in how the parts were put together. And it rarely sits inside one department, which is why nobody owns it. A few patterns show up again and again:
It works in the demo. Then it meets a strange input, a load spike, or a compliance question no one planned for. I build for the ordinary Tuesday, when the thing just has to hold.
Point AI, or any tool, at a broken workflow and you get broken output. Faster. I fix how the work actually flows first, then automate what's left.
Shipping is cheap now. Deciding what's worth shipping is the hard part, and it's where the money quietly leaks. Some of the best work I do is talking a team out of a build they'd already started.
What usually breaks after someone like me leaves is a handover that never really happened. So I build it so your team can carry it, write down what I did, and get out of the way.
I find the places in your business where AI genuinely takes work off people, and the places where it just adds a shiny new liability. Then I build the first kind. At AXB that was 5+ applications live in four months, across compliance, reporting, and trade analytics. Execution latency cut tenfold.
I open up what you already run (code, infrastructure, the AI stack) and give you a ranked, plain read: what's solid, what's fragile, what to fix first. Data first, then the flows, then the models on top, because getting that order backwards is why most of it falls over.
The unglamorous half of the job, and usually the half that matters. I map how work really moves through your company, cut the steps that survive out of habit, and hand the rest to AI once it's genuinely worth handing over.
Build it, buy it, or leave it. I've made these calls across trading systems, an exchange handling thousands of requests a second, and a venture builder behind 42+ companies. So you get judgment, not a shrug. Including the times the honest answer is "you don't need this yet."
There's no shortage of people who'll sell you a diagnosis, and plenty more who'll build whatever they're handed, right or wrong. I do both.
I've been the engineer, the architect, the CTO and the CEO. I've owned the code, the roadmap, the P&L and the board conversation, so I can read a problem from all of them at once. I work out where the problem really sits, then I'm the one who fixes it. Same head on the diagnosis and the delivery, so nothing falls down the gap between them.
Fixed prices where they can be fixed. A straight conversation where they can't. No games, no request-a-quote runaround.
You point at what's wrong. Two weeks inside your company to understand it properly. Then, depending on what it is: I fix it, hand you the plan to fix it, or we start a longer piece of work. The number is fixed before we start, and credited in full against a Build or Retainer within 60 days.
Book the auditWhen you know what needs to happen and want it built properly. AI in production, a systems rebuild, a process overhauled and automated. Scoped up front, milestones agreed before anything starts, built with your team so it stays yours.
Scope a build →A standing senior pair of eyes across the whole company. I stay close to the systems, make the big calls as they land, review what's shipping, keep the whole thing honest. Not managing your people. Looking after the ground they build on. Pricing is a conversation: tell me the situation and I'll be straight about it.
Talk retainer →A ranked read on what's actually holding you back, and what to do about it first
AI in production and earning, not stranded in a pilot
Processes that flow, then automate, not automated chaos
A clear make-buy-or-skip call on the things you were about to fund
Fewer things breaking at 2am, because the architecture got looked at before it shipped
A company your own team can keep running the day after I'm gone
Point me at a problem and give me room to understand it. That's the whole ask. Understanding is most of the work. That's why the front door is a two-week audit, not a slide deck.
I'll mentor your people, but I won't run them. I'm not a stand-in manager. I've built companies of 160+, so I'll happily bring your senior people up a level along the way. But your org chart stays yours.
I build it to work without me. Every engagement is aimed at the day I leave. Everything gets documented, the framework underneath is open, and your team owns what's left, not a black box only I understand.
I'll say the awkward thing. Where the money's going, which pet project to kill, why that "quick" fix is three months of work. That candour is most of what you're paying for.
Two clients at a time, hard cap. I'm not going anywhere. AXB is my main bet. The cap is what makes one or two days a week real attention instead of a sliver.
Backend engineer, architect, CTO, CEO: the whole climb from the code to the boardroom, across fintech, crypto and trading, always in regulated, high-stakes places. The thread through all of it: turning tangled, manual processes into systems that scale, without burying them in process.
Rebuilt a live digital-asset and FX trading firm around AI: 5+ applications in production in under four months, execution latency down 10x, a team running on agentic tooling every day. Trading live across 10+ venues.
Architected one of Europe's largest crypto exchanges: a million-plus users, a platform handling thousands of requests a second. Built a trading module that cleared ~€250K in a single day.
A venture builder that's launched 42+ tech ventures, several of them licensed financial institutions, with 160+ people worldwide. I ran the innovation lab and set the bar for what got built: 500+ ideas assessed, and I killed far more than I shipped.
The AI framework I actually build with, published in the open. The way I build isn't a secret; it's on the table.
And 10+ other ventures across fintech, crypto and digital infrastructure.
I'm not a full-time hire, and at this stage you probably don't need one. You need someone who's built this before and can point your team at the right thing. I'm that person.
No. I won't become your manager. I'll mentor your senior people and work alongside them, but the team stays yours. Done right, the people you have get sharper, not replaced.
Both. AI's usually in the mix, because that's where a lot of the quick wins sit right now. But the real job is finding what's stuck (architecture, process, product, cost), and AI is one tool for that, not the point of it.
Consultancies hand over a strategy and leave. Agencies build whatever the brief says. I work out what's worth doing and build it with you. One person you actually deal with, not a rotating cast.
I'm not going anywhere. AXB is my main bet. That's exactly why the cap is two clients: this works because it's small, senior, and scoped, not because I'm looking for a full-time seat.
Usually within a week or two of a call.
Based in Poland, working remotely wherever timezones allow. Mostly founders and leadership teams at fintechs and scaleups.
You own all of it. Everything's written down, the framework underneath is open, and your team keeps going without me. That's the point of doing it this way.
A short call. Tell me what you're building and where it's stuck, and I'll tell you straight whether I'm the right person for it, and whether you even need me yet.