Senior engineering leadership for teams building difficult digital products.
I'm Stéphan Zych, an engineer and technical lead in Brussels. I'm brought in to clarify, lead, ship, or rescue a project that has gone sideways. One senior partner, from the first call to production.
Brussels-based · Remote-ready · Taking on a small number of advisory, leadership, and delivery engagements
problems
The problems I'm brought in to solve.
Not a list of services: the situations that actually get a senior engineer a call. If one of these sounds familiar, the next move is a conversation.
legacy
The platform works. Nobody trusts changing it.
Every change feels risky, so the team slows down to avoid breaking something nobody fully understands. I clarify the system, reduce the risk of touching it, and ship improvements without pausing the business.
delivery
There's a roadmap, but nothing ships reliably.
The plan exists; the delivery doesn't. More features will not fix an unclear delivery process. I find where work actually stalls, tighten how things get built and shipped, and get reliable progress back.
scale
The product grew faster than its foundations.
What got you here is now the thing slowing you down. I evolve the architecture so it can carry the next stage, without a full-stop rewrite or a six-month freeze.
direction
You need senior technical direction, not a full-time CTO.
The decisions are real, but a permanent hire is too much. I provide senior direction and decision-making at the level you actually need, for as long as it's useful.
partner
Your agency needs a trusted lead engineer for a hard build.
A complex engagement needs someone senior who can own the technical side and represent it well to your client. I work as an extension of your team, not a black box.
rewrite
A rewrite is on the table. Nobody has assessed the real problem.
A rewrite is not always the answer. Sometimes the answer is finally understanding what exists. I assess the system honestly and tell you whether to rebuild, refactor, or leave it alone.
ai-guardrails
Everyone here uses coding agents. Nobody agreed on how.
The tooling is already in your codebase, but every developer runs a different setup, the output gets reviewed to no consistent standard, and nobody owns the guardrails. I move the configuration into one shared, versioned source, make reviewed diffs and deterministic checks the condition for merging, and sandbox what the agents are allowed to reach.
capabilities
What STEPHANZYCH actually does.
A small number of things, done at a senior level: from technical direction down to the code itself.
leadership
Technical leadership & delivery
For
Teams that need senior direction, stronger technical decisions, and reliable progress.
Work
Set technical direction, make the hard architectural calls, unblock delivery, and stay close enough to the code to keep the decisions honest.
Outcome
Decisions get made, work ships on a cadence people trust, and the team carries the standard after I leave.
architecture
Architecture & platform evolution
For
Existing systems that need to become easier to understand, change, operate, and scale.
Work
Map what's really there, expose the constraints, and evolve the architecture in shippable steps, not a big-bang rewrite.
Outcome
A system that's cheaper to change, safer to operate, and ready for the next stage of growth.
product
Product engineering
For
Teams building substantial digital products, from early technical direction through production.
Work
Own the build end to end (interface, application, and the APIs and infrastructure underneath) with a plan up front and a working product at the end.
Outcome
A real product in production, documented and owned by you, not a prototype that needs rescuing.
audit
Audit, advisory & due diligence
For
Organisations that need an independent technical view before an expensive decision.
Work
Assess the codebase, architecture, team, or proposed plan, and report findings and recommendations you can act on straight away.
Outcome
A clear, defensible read on risk and a recommendation: extend, rebuild, invest, or walk away.
enablement
Team enablement
For
Teams that want stronger engineering practice, workflow, and technical ownership.
Work
Improve how the team builds (code quality, review, tooling, and delivery workflow) through hands-on work on their own codebase, not slideware.
Outcome
A team that ships with more confidence and needs less rescuing over time.
ai-engineering
AI-assisted engineering
For
Teams already using coding agents, where every developer has a different setup and nobody owns the guardrails.
Work
Move the agent configuration off individual machines into one shared, versioned source that applies whichever tool a developer drives, put deterministic checks and reviewed diffs around every change, keep sandboxing outside the tooling's control, and say plainly where this pays and where it does not.
Outcome
Agent-assisted work that is reviewable, reproducible across the team, and bounded by rules the tooling cannot switch off.
Same senior person, three commercial shapes. It comes down to two questions: do you need me ongoing or for a defined piece of work, and do you need something built or assessed? The first conversation works out the fit. It doesn't force a package.
Embedded technical leadershipOngoing · retainer or set days per week
Focused project deliveryFixed scope · defined timeline
Audit, advisory & due diligenceFixed scope · days to a few weeks
Leading full-stack delivery and technical direction on enterprise platforms, engaged through STEPHANZYCH.
Cherry Pulp2013 — 2024Co-founder & CTO
Built the studio and its engineering team from zero to acquisition.
TBWA · Agency.com2009 — 2013Rich-media Developer
Interactive and campaign work for international brands.
0 → exit
From zero to acquisition
Cherry Pulp
A studio that grew from zero to a successful acquisition in 2023, with the transition completing in 2024 and an engineering capability that outlasted any single project.
Modernisation in progress
Modernising how a team builds and ships
CBTW
In progress. The shared local environment is in place, and the pipeline is rolling out project by project alongside ongoing client delivery: legacy workflows and the new one currently run side by side.
Held under COVID peaks
Re-platforming a food business under COVID load
eFarmz
The new platform carried eFarmz through the unprecedented order peaks of the COVID crisis with real-time ERP sync intact. Work that contributed to an E-commerce of the Year award and recognition at the BeCommerce and Feweb awards.
Scaled with the traffic
Scaling an automated magazine platform
MyTribuNews
Magazines kept generating within hours as usage grew exponentially, and the platform held performance and availability through its peaks.
Landmark-scale clients
Making district-cooling data legible
Climespace
Building owners got a single place to manage cooling budgets, consumption, and equipment, with their installations rendered as a clear schematic rather than a spreadsheet.
Tens of thousands of runners
The city's race, from sign-up to finish
20km de Bruxelles
A trilingual home that carried runners from sign-up to finish line (registration, practical info, and results in one place) for one of Belgium's largest races.
Four brands, one platform
Four brand sites on one shared theme
Daoust
Four Daoust brands now share one design and one platform, consistent to visitors and far lighter to maintain than four separate sites.
Full SaaS frontend
Self-storage with a customer portal
Boxify
Boxify customers got a portal they can size, fill and manage their storage from, and the team got the admin tooling and API to run the service behind it.
One codebase, two platforms
A mobile app for a historic market site
Abattoir
Abattoir got one cross-platform app on a single codebase and API, rather than two native builds to keep in sync.
Mon Climespace won the Public Prize at ENGIE France Réseau's Innovation Trophies — voted by staff who loved its content, features, and usability. Wonderful collaborative work across the CLIMESPACE, Cherry Pulp, ENGIE IT and Bonne Pioche teams throughout this award-winning project.
We were able to keep our deadline and produce a quality website that has already recorded new automatic client on-boarding.
Jean-Louis Van HouweFounder & CEO, Monizze
I couldn't leave the office without thanking you and your team for your flexibility and availability. As I often say, it's a genuine pleasure to work with you and to be able to count on a fantastic team.
Oscar ChavezBusiness Europe
The "20km of Brussels" are our child — and you understood that perfectly. Thank you for our new website: it's elegant, radiates simplicity, and shines.
Carine Verstraeten20km de Bruxelles
operating-principles
How I think about the work.
Understand before replacing.
A rewrite is not a strategy. Sometimes it's necessary; more often it's a way to postpone understanding the system that already exists. The assessment comes first.
Architecture is a delivery tool.
Good architecture should make change cheaper, not make diagrams prettier. If it isn't helping the team ship, it isn't doing its job.
Ship improvements while reducing risk.
Progress and safety aren't a trade-off. The goal is to make things better without pausing the business or betting it on a big-bang release.
Make complexity visible.
The hard part is rarely choosing a framework. It's seeing clearly what's actually there. Naming the real constraints is most of the work.
Use the simplest system that survives reality.
Not the cleverest, not the trendiest: the one a real team can operate, change, and reason about under pressure.
Seniority means owning the trade-offs.
Senior engineering is not adding meetings around a problem. It's taking responsibility for the decisions and their consequences.
Documentation is part of the product.
Work that only one person understands isn't finished. What I deliver is meant to be operated and changed by your team, not just by me.
Leave the team stronger.
The best outcome of an engagement is a team that needs me less afterwards, not more.
about
Fifteen years of shipping, available as a partner.
Stéphan Zych is a senior engineer and technical lead who works the whole stack across agency, product, and enterprise environments. Co-founder and CTO of Cherry Pulp through its acquisition; now lead developer at CBTW, engaged through STEPHANZYCH.
The work sits between business needs, delivery pressure, and engineering quality: taking ownership of complicated problems, making the calls, and implementing the difficult parts. The more experimental, personal side of the work lives at stephan.zych.be.
Founder & CTObuilt and led an engineering team to acquisition
Lead developeron enterprise web platforms
Full rangearchitecture, delivery, leadership, and hands-on implementation
Brussels · remoteavailable across Belgium and Europe
faq
Questions, answered.
What kind of projects do you take on?
Difficult ones: complex web platforms, delivery that's stalled, architecture that needs to evolve, and decisions that hinge on the real state of the code. Less interesting: generic builds that any team could do.
Do you work remotely or on-site?
Remote by default, from Brussels, and set up for it. On-site across Belgium and nearby is no problem for kickoffs, workshops, or when a team simply works better in the room.
Can you join an existing team?
Yes. That's a lot of the work. I embed as a senior engineer or technical lead, work the way your team works, and represent the technical side well to stakeholders and clients.
Can you help with a legacy platform?
Yes. Clarifying and evolving systems that have grown complicated is core to what I do. Usually the first job is understanding what's really there before changing anything.
Do you only advise, or also implement?
Both. The advice is more useful because I still write the difficult parts. I take responsibility for decisions and then help make them real.
Can you work with agencies and their clients?
Yes. I act as a trusted lead engineer on complex engagements: an extension of your team that you can put in front of a client, not a black box.
Can you assess a proposed rewrite or architecture decision?
Yes. A fixed-scope assessment that tells you honestly whether to rebuild, refactor, or leave it alone, before you spend the budget.
Fixed scope, retainer, or day rate?
All three, depending on the problem. Project work is usually fixed-scope; leadership and advisory work runs on a retainer or day rate. Agreed up front, in writing.
Do you use AI tooling on our code?
Yes, deliberately and under constraints. Agent work goes through reviewed diffs and deterministic checks like any other change, runs sandboxed with limits the tooling can't lift, and the rules live with your project rather than on my machine. You decide what it may touch and what never leaves your environment; if the answer is local-only, or not at all, that's the setup.
Who owns the code and IP?
You do. Everything is delivered to you, documented, and yours to keep. I'm glad to sign an NDA before you share anything sensitive.
How do you start an engagement?
A short call to understand the situation, then a clear proposal: scope, approach, and price, or an honest "not the right fit." Work starts against a plan you can hold me to.
start
The next move, not another roadmap.
Tell me what's stuck. The first call is to work out whether I can be useful, not to sell you anything.