AI is a tool, not a replacement for thinking. The most effective AI-assisted development combines the speed of large language models with structured processes, clear documentation, and human oversight. Context matters—the quality of your output depends on the quality of your guidance. And engineering fundamentals never go out of style.

Most AI projects fail because they skip the fundamentals—proper architecture, clear requirements, testing, documentation. The technology usually isn’t the problem. It’s treating AI like magic instead of engineering.

Claude (Anthropic), OpenAI GPT models, and the ecosystem around them. I focus on practical implementation—APIs, orchestration, guardrails, observability—rather than model training or research.

Not every problem needs an AI solution. Sometimes the right answer is a well-designed database query or a simple script. I help organizations identify where AI adds genuine value versus where traditional engineering approaches are more appropriate, cost-effective, and maintainable.

My path wasn’t typical—I started as an intensive care nurse, studied political science, then transitioned to software engineering. Healthcare got me interested in data (ICU patients generate mountains of it), and political science deepened that with statistics and research methods. That data work eventually pulled me back to a hobby I’d started as a teenager in the early 90s: programming. Healthcare taught me that systems must be reliable under pressure. Political science sharpened my understanding of organizational dynamics. This cross-domain thinking helps me bridge technical implementation and real-world impact.

Various projects focused on AI integration and data for Dutch ministries (Finance, Justice & Security, Health, Tweede Kamer) and enterprises like ArboUnie and Acta Marine. Notable examples include Beter Bereikbaar Applicatie (won a Smart City Award) and Meldapp (smart building management platform).

Primarily Dutch government (ministries, municipalities) and healthcare organizations. I’ve also worked with maritime, construction, and general enterprise clients. The common thread is complex requirements and high stakes for reliability.

I work as Lead Developer at ePublic Solutions, but I’m available for freelance and consulting work. Typical engagements include architecture reviews (1-2 weeks), technical advisory (ongoing), proof-of-concept development (2-4 weeks), and full project implementation. I work remotely across Europe, with occasional on-site visits for Netherlands-based clients. Reach out via the contact button below to discuss your specific needs.

It starts with understanding your actual problem—not the technology you think you need. From there, it depends: maybe it’s a week-long architecture review with documentation and recommendations. Maybe it’s building a proof-of-concept to validate an approach. Or a multi-month engagement building production systems with your team. Every project includes clear documentation, knowledge transfer, and a focus on maintainability.