About Jan Dekkers
I’m Jan Dekkers, a fractional CTO, technical advisor and software architect with thirty years of shipping production software. I write AI Web Center: practical, plain-language guides to adopting AI in software work, especially AI-assisted programming.
Background
I’ve spent three decades architecting and shipping production software across regulated healthcare and diagnostics, financial services and embedded systems, from firmware and custom hardware to enterprise platforms and cloud integration. My specialty is inheriting failing legacy architecture and turning it into something durable.
Some highlights:
- Financial services. For 17 years I was lead software engineer at Ascensus, reporting directly to the CFO and owning a six-application platform for retirement plan administration that served 1,000+ users across four offices. Along the way I digitized about 7 million pages with OCR and cut 8 FTE days a month of manual billing work.
- Clinical diagnostics. At SMedix, contracted to Roche, I integrated 2D barcode handling into Ventana cancer-staining software, saving Roche $6M a year and earning the Tom Grogan Award, and built a bridge between two incompatible generations of Ventana BenchMark ULTRA instruments.
- Imaging and embedded systems. I’ve built DICOM medical-imaging conversions, award-winning imaging components, a contactless health-monitoring device for MRTX, and my own patent-pending digital position readout for lathes.
Today I work as a software engineer at a US software-as-a-service company, remotely from Huntington Beach, California, and I advise companies as a fractional CTO. See my resume for the full story.
How I use AI
I’m technical lead on an AI-assisted ground-up rewrite of a production software platform, delivered by a very small team, with the large majority of the production code generated by AI and every line of it reviewed.
What I’ve learned: the bottleneck isn’t the model. It’s specification, decomposition, and judgment about what the generated system will cost to own. That’s why the guides here put as much weight on rules files, small steps and review as on prompting.
Why I started AI Web Center
Most writing about AI coding is either vendor marketing or predictions about developers being replaced. AI Web Center is the practical counterpart: what the tools actually do today, how to set them up, and how to check their work, written for developers and for the business owners who have to decide whether to adopt them.
How I write
- Plain language. If a term needs explaining, it goes in the glossary.
- Verified facts. Claims about tools link to the vendor’s own documentation, with the date they were checked.
- No hype. I describe what tools do today, including their limits.
- Corrections welcome. If something is wrong or out of date, email me and I’ll fix it.
Work with me
I advise on AI-accelerated delivery, legacy modernization, architecture reviews and technical due diligence, and regulated and medical software. Email hello@aiwebcenter.com or use the contact form.