I'm a senior software engineer with 12+ years in production systems. I work inside development teams in Belgium and help each team make AI part of how it builds software.
Most IT companies bought licences. Far fewer changed how they work, so the gains stay small and uneven.
Every developer uses AI in their own way. Nothing is shared, so nothing compounds.
Nobody has agreed what AI may touch, how AI-written code gets reviewed, or where client code is allowed to go.
Without a baseline, you can't tell whether AI saves time or just moves the work to code review.
Project-based, on-site or remote
If your team needs another senior engineer and its AI tools aren't paying off yet, I can take on both. I join the team and work on the code like everyone else. While I'm there, I set up how the team works with AI: shared conventions, review rules for AI-written code, and worked examples on your own tickets.
You get
The work delivered, and a team that keeps working this way after I leave.
Typical projects
The conventions I help a team agree on look like this: what AI may draft, what needs a second reviewer, and how the team measures the effect. Feel free to copy it.
# team-ai-conventions.md
## AI may draft
- tests, migrations, docs, boilerplate
- refactors, only where test coverage exists
## Needs a second human reviewer
- auth, payments, secrets, infrastructure
## Rules
- AI-written PRs meet the same review bar; label them `ai-assisted`
- client code only in tools with zero data retention enabled
- measure cycle time and review time monthlyYou send a short brief: your team, your stack and what you need. If it looks like a fit, we talk for 30 minutes. If it doesn't, I'll tell you.
A short written offer: price, dates and what's included.
One team and one project to begin with, and we agree on what to measure.
What works gets rolled out to other teams, documented and taught. What doesn't gets dropped.
Client names are confidential. Details on request.
AI coding agents, Java, Kotlin
Developer productivity, Nx, React
Architecture, .NET, Azure, gRPC
Training, architecture
AI integration
The human in the loop
Senior software engineer, 12+ years in production systems
I'm Laur, based in Leuven, Belgium. For 12+ years I've built and run production systems, from distributed backends and web platforms to cloud migrations, and led teams through large rewrites. Along the way: Kafka messaging, micro-frontends, CI/CD pipelines in Azure and a reporting service on Apache Superset.
Today I use AI in every part of that work. My engineering background is the point: I know when AI makes a team faster, and when it just produces more code to review.
I hold a Master's in Distributed Systems and work in English.
I write code. Most of my work is as a senior software engineer inside a team, and I bring the AI workflow into that team while I work.
Mainly Claude Code, GitHub Copilot, Cursor, and the OpenAI and Anthropic APIs. I recommend what fits your stack, budget and security requirements, whichever vendor that is.
Before anyone uses AI on your code, we agree which tools are allowed, with which data-retention settings, and how AI-written code is reviewed.
No. I'm most useful to IT companies with roughly 5 to 100 developers, and I can start with a single team.
Yes, on-site in Belgium, or remotely.
Send a short brief: your team, your stack and what you need. I reply within two business days, and if it looks like a fit, we set up a call.