Belgium · Independent

Software that still makes sense six months after launch.

I work shoulder-to-shoulder with founders and product people: AI where it helps, custom code where off-the-shelf won’t cut it, web apps when the interface is the product. You get straight answers, short feedback loops, and no mystery team behind the curtain.

What usually goes wrong

Rarely “not enough developers.” More often: fuzzy scope, integrations that break under load, or software that doesn’t match how people really work.

Figure the problem out first

I sketch flows, constraints, and risks before touching production code. Stops us from coding the wrong thing really efficiently.

You talk to me

No account manager in between. Quick decisions, fewer meetings, and you always know who’s on the hook.

Built to last past the demo

Structure and tests where they earn their keep, so when traffic, data, or compliance pressure shows up, the system doesn’t fall apart.

What that can look like

Numbers from real-style projects (illustrative; your mileage varies). The point is milestones, ownership, and something you can measure.

Back-office automation
−82% time on manual processing

Intake and validation pipeline for a logistics operator: less copy-paste, fewer errors.

Customer-facing app
4.2× faster releases

Modular web app with CI, basic observability, and rollouts you can actually control.

Decision support
−35% time to an answer

Structured data plus LLM workflows for a B2B services company: not magic, just less digging.

Where I spend my time

Three buckets. Same deal each time: ship something useful, write down the important calls, leave code someone else could pick up.

All services

AI systems

Automation, search over your own data, evaluation, guardrails, wired into how you already work.

Custom software

Internal tools, APIs, integrations, domain modeling when spreadsheets and SaaS templates stop scaling.

Web applications

Interfaces people actually use: performance, auth, and edge cases handled, not just the happy path.

How a project runs

01

Discover

Agree what “good” looks like, map constraints, pick the smallest architecture that could work.

02

Build

Slices you can see and test: demos, tests where they matter, notes so your team isn’t left guessing.

03

Scale

Hardening, logging, iteration when real usage shows you what hurts.

Why one person isn’t a downside

You always know who’s building. I’m not juggling five accounts or handing you off to a junior with my name on the proposal. Trade-offs and bad news come from me directly. If you later need a bigger team, I’ll help you hire or onboard without trapping you in my stack forever.

Got a bottleneck or a half-baked idea?

A few sentences is enough. I’ll tell you honestly if I can help, or point you somewhere else.

FAQ

Do you sit in with our team?

If it helps, yes. I’ll join standups or planning sessions when that speeds things up. I still own delivery on agreed milestones.

How big are your projects?

Often chunks of four to eight weeks, sometimes longer programs. I price and plan around outcomes, not an open-ended hourly tap.

Can you plug into what we already run?

That’s normal work. APIs, databases, SSO, weird legacy exports. I assume your stack exists and plan around it.