Data architect with 20+ years in ICT and a hybrid technical + business profile. I design auditable, change-resilient enterprise data warehouses — and explore how agentic AI can remove the repetitive parts of that work without removing the architect’s judgement.
Focus
- Data Vault 2.0 / 2.1 — modeling and automation (CDVP², 2015; 2.1 refresh in progress)
- Agentic AI for data engineering — LangGraph, Anthropic Claude, AutomateDV / dbt
- Requirements engineering, data contracts, DACH enterprise data platforms
Selected work
- Vault-Agent — open-source agentic Data Vault 2.0 automation: requirements → model → contract-backed dbt code, with a transparent decision trail.
Background
Over 20+ years in ICT I’ve delivered data-warehouse, BI, regulatory-reporting and data-quality programmes — mostly for Swiss banks and insurers, with detours through healthcare, telecom and retail. The work spans the full chain from requirements and data modelling to ETL and acceptance testing, and the constant has always been the same: turning messy business and regulatory needs into something auditable and maintainable.
Since 2011 I’ve worked as an independent consultant for Swiss tier-1 and cantonal banks, large insurers, a national retail-banking provider, and major telecom and healthcare players. Before going independent I spent seven years at Teradata Switzerland building enterprise warehouses for some of the country’s largest telecom and retail organisations. These days I’m just as interested in where agentic AI fits into that craft — which is what this site is mostly about.
I model in Data Vault 2.0 (CDVP², 2015) and work in requirements engineering, mostly around Oracle, Informatica, Teradata and SQL Server, with SAFe and Scrum where it helps. Fuller detail lives on LinkedIn; the short version is that I like data platforms that are still understandable a year after they’re built.