Career history
In the news
- You've heard it a hundred times: AI is only as good as the data behind it. What you may not have heard: on Databricks, you can fix your data quality without copying your data. It often works like this: first a copy goes to the database of your DQ tool. That means a second database, a second set of access rights and a copy that is always behind. With DATAshaper pushdown we do it the other way around. The rules go to the data. The checks run as SQL on your own Databricks SQL warehouse. The source tables stay read-only. The issues
- Amplr op bezoek bij tillit en DATAshaper in Lokeren 🛜 Tillit is een IT-consultancybureau, gespecialiseerd in data. Met hun eigen tool DATAshaper helpen ze bedrijven met datamigraties. Vroeger viel hun internet geregeld uit waardoor een deel van het team, of zelfs iedereen, van thuis moest werken. Dat hebben we met Amplr opgelost. En nu kijken we samen naar NIS2. Merci Kevin Sermijn en het hele team om ons te ontvangen!
- Our internal AI hackathon was about timesheet accuracy. A textbook example of data quality as the foundation for AI. At tillit and DATAshaper we believe in practice what you preach. So we tackle our own processes too, not just those of clients. Timesheets feed invoicing, planning and reporting. Without correct data going in, no reliable insights or AI coming out. Five teams, one problem, each from a different angle. Our framework was the common thread: - Strategy: fewer wrong invoices and credit notes, so direct time
- Sovereign AI, applied to not the sexiest topic: an ERP data migration. We needed the Installed Base, the equipment installed at our customers. It didn't exist as structured data anywhere. It was hidden in thousands of free-text customer notes. The classic approach: a team reading note by note for days and retyping serial numbers, brands and models into an Excel. Repetitive, slow and extremely boring work. Instead we let an LLM read the notes and extract everything into a structured dataset that DATAshaper prepares for SAP.
This page shows public professional information only, each fact cited. Is this you? send a correction, or ask for removal within 24 hours, no questions asked.