Senior engineering manager · Amsterdam area · Open to hybrid or remote
I build the platform, then grow the people who run it.
Engineering manager, and a software engineer for seven years before that. In 2018 I built a company’s first analytics platform. I have led it ever since, through three acquisitions, to around 900 million events a day.
Along the way I grew the team behind it from five people to ten, grew its leads and trained my successor. Then I took the same approach to a three-team platform migration. Every senior and lead on my teams was grown in-house.
Today I lead 8 engineers in two teams. At the 2024 peak it was 11, including a team lead and her team, and the migration programme I led spanned 13 engineers in three teams.
What changed on my teams.
every 2–3 months→every sprint
Release cadence, after automated end-to-end tests and releases
80% maintenance→75% projects
My data team’s time, eight months after I unblocked a stalled handover
12 hours→under 2
A full cycle of ~400 Spark jobs, after a custom scheduler
up to 2 min→1.5–10 s
Report load time, after rebuilding the reporting API on Trino
01 — Selected work
Three stories, each with its trade-off.
What the situation was, what I decided, what it cost, and what changed.
~900Mevents a day, on a platform I built with one colleague in 2018
The platform I built, and the team I built around it
Situation
The company had no analytics. Customers would get it free with the product, so it had to run at close to zero cost.
What I did
Built the first version with one colleague in three months, and we became the first team in the company on Kubernetes.
Redesigned data collection so customers integrated once. Every later improvement happened on our side, which made the product much easier to sell.
As team lead, moved event storage from Redis to S3, which halved the running cost, and built a scheduler that took a full cycle of ~400 Spark jobs from 12 hours to under 2.
As manager, rebuilt the reporting API on Trino and delivered a full redesign of the user interface.
The trade-off
In 2026 I paused feature work on my own platform so its team could join the Kubernetes migration, which the business needed more.
What changed
25 microservices handling ~900M events a day, used by almost 90% of the platform’s customers. I coached a senior QA into the lead role, and she was running the team on her own within six months.
3 → 13engineers, as the programme grew from one team to three
Cross-team leadershipSearch platform migrationCrownpeak, now Rezolve AI2024 – 2026
Led a three-team move from EC2 to Kubernetes, largely through people outside my line
Situation
The search platform ran on EC2. The aim was to move it to Kubernetes and cut running costs by around 30%. It started as a three-person project covering a third of the platform.
What I did
Grew it into a programme of three teams and 13 engineers covering ETL, search and navigation, orchestration and suggest.
As product owner for the whole programme, set the delivery commitments and every team’s milestones, led the third team through its team lead, and made the technical calls through most of the build.
Rebuilt the catalogue ETL, the part everything else depended on, on Apache Hop and Argo Workflows. I set the architecture and coached two data engineers who had never written production software through the build.
Alongside it, redesigned the department’s team structure with the Head of Engineering: two teams split into three, with clear boundaries and a common way of working. Release cycles got shorter on some components, maintenance and support tickets were answered faster, backlogs became more focused, and team leads spent less time on admin.
The trade-off
I kept work in progress in line with team size by stopping the least important projects, rather than spreading people thin.
What changed
The programme met its milestones, and the ETL went to production on schedule, with pipelines up to 80% faster and up to 50% cheaper.
Two teams reported to me; the third I led through its team lead.
Me · product ownercommitments, milestones, technical arbitration
Teams 1 and 23 developers each · my teams, reporting to me
Team 37 engineers including its team lead · led through that lead, outside my line
+ 2–3 infrastructure engineerssupporting all three teams
reports to meled through its team lead
6 monthsfor the senior QA I coached into my successor to run the team on her own
PeopleInsights and data teamsCrownpeak → Rezolve AI2020 – 2026
Growing people into owners
Situation
Small teams, no recruiter, no levelling framework, and a product that had to run on almost nothing.
What I did
Hired two juniors myself: wrote the posting and the technical assignment, screened more than 100 CVs and ran the interviews. Both were promoted within a year.
Turned an under-performer around with an honest conversation instead of a formal improvement plan, then gave them ownership of training the new juniors. A year later they were promoted to senior.
Grew two team leads: a QA promoted into a new lead role after the reorganisation, and a senior QA coached into my successor.
Set the promotion bar myself. With no framework, I built every promotion case on evidence.
Designed out single points of failure: a pair-programming rotation, epic owners, a production-support rotation, and no more than two projects at a time.
What changed
Every senior and lead on my teams was grown in-house. When engineers later left, we lost capacity but no knowledge.
“I’ve watched people on your team go from asking permission to just doing things.”
02 — How I lead
Five habits, each with its proof.
Hire for attitude, then hand over real ownership.The two juniors I hired were promoted within a year, and the under-performer I kept went on to train them.
No single points of failure.Rotations, epic owners, two projects at most. When people left, the knowledge stayed.
Fewer things at once.I stop the least important project before I spread people thin.
Choose what the team can run for five years, and think about cost first.A free product that had to run at close to zero cost, and half the running cost after the move to S3.
Make the reasoning visible.Direction is agreed with the team, not handed down. The AI guardrails were agreed with the people who use them.
“A rare combination of technical depth and the drive to bring others along for the journey… He makes the reasoning behind architectural and product decisions visible to everyone, not just those in the room.”
“You’re good at staying calm in chaos and bringing people with you… You pick the thing the team can run for the next five years. You think about cost before anyone asks.”
“You still do the boring migrations yourself, so people listen when you ask them to do theirs.”
03 — How I work now
AI-assisted engineeringat work in 2026 homelab since Sep 2024
Two teams moved in two months, and a homelab run through agents
At work
I moved two teams from occasional Copilot use to spec-driven development with Claude in two months. I set the direction and agreed the guardrails with the teams: tests first, AI review before human review, and a person signs off every spec and plan. The teams built the tooling. The 8 engineers delivered about 20% more tickets and features, and quality went up, because the extra capacity went into maintenance and CI/CD.
At home
A 7-node Kubernetes homelab running ~70 services, run day to day through Claude Code. About 400 of its 1,550 commits were AI-assisted. The agents work under guardrails I designed, with four levels of autonomy, hard limits no agent can change, and rules enforced in code rather than in prompts.
A decision
I built a 15-role autonomous agent fleet, then retired it when its running cost outweighed its value. The guardrails outlived it.
The homelab is where I try things out before I bring them to work. The tooling my teams use at work is theirs.
04 — Career
Four company names, one job.
Ten years on the same e-commerce search platform while it changed hands three times, after three years at Esker in Lyon. Each role carried over unchanged through the acquisitions.
Apr 2023 – present
Software Development Manager
Rezolve AI, Amsterdam · Nasdaq-listed, about 1,400 people · acquired Crownpeak in Dec 2025
Two teams on an e-commerce search and merchandising platform, one of them owning the analytics platform I founded. Reporting to the Head of Engineering.
Led the EC2 → Kubernetes migration across three teams (story).
Redesigned the department’s team structure with the Head of Engineering, and promoted a QA into one of the new lead roles (story).
Moved two teams to AI-assisted development in two months (below).
Freed my data team from 80% maintenance to 75% project work within eight months.
Crownpeak, Amsterdam · about 350 people · acquired Attraqt in Dec 2022
Led the analytics platform team: grew it from five people to ten direct reports, and took on a second team of three .NET developers, whom I moved onto our Java and Scala stack within a year.
Halved the platform’s running cost and cut the job cycle from 12 hours to under 2 (story).
Took releases from every two or three months to every sprint, with no manual test cycle.
Made my team the first in the company to own its monitoring. Failures that once went unnoticed for up to two weeks were fixed by the next day.
Hired two juniors, turned an under-performer around and set the promotion bar (story).
Leadership
Hiring and onboarding, performance management, promotions, team design, coaching