About Me
AI/ML Initiative Lead & Enterprise Architect
Based in Muri bei Bern, Switzerland. Over 25 years building large-scale distributed systems at the intersection of software architecture, machine learning, and agentic AI. Currently leading AI initiatives at SBB.
AI & Agentic Systems
Designing domain-expert agents coordinated via Model Context Protocol (MCP). Exploring formal verification, schema-driven interoperability, and LLM tool scaling.
2025–PresentLarge-Scale Architecture
Led architecture for SBB's omni-channel distribution and next-generation sbb.ch platform. Enterprise AI system design.
2013–PresentMachine Learning
Statistical learning (HMM, GMM, SVM) for handwriting recognition, writer identification, and document image analysis.
2004–2007Formal Methods
Process algebra (π-calculus), meta-modeling (MOF/XMI), component-oriented programming. 20-year arc to agentic verification.
2001–2003Education
PhD in Computer Science (summa cum laude)
University of Bern, 2008
Writer Identification and Verification
Diplom (Master's)
University of Bern, 2002
Enabling White-Box Reuse in a Pure Composition Language (JPiccola)
Studies in Philosophy & Psychology
University of Bern, alongside CS degree
Philosophy with focus on logic and philosophy of language, logic and ethics; psychology with focus on cognitive psychology and philosophy of mind. Cognitive science has been a lifelong interest, and the interplay between formal logic and empirical psychology continues to shape how I think about intelligence, both human and artificial.
Open Source & Standardization
Long-standing contributor to open-source projects including KDE KHTML and WebKit CSS parser (2001–2005).
Technical Lead & Product Owner of the Open Sales and Distribution Model (OSDM) — an international REST API standard for rail ticket distribution. Adopted by all major European railways (SBB, DB, SNCF, ÖBB) and leading global distribution systems.
Interests & Philosophy
I'm drawn to questions at the boundary of formal specification and running systems: when does a schema become a contract, and when does a contract become a proof?
More over I'm deeply interested in programming language design, especially their semantics and expressiveness for modeling complex systems. Especially im interested in type systems, and how they can be used to capture invariants and enable safe composition in large-scale software.
Furthermore, I'm deeply fascinated by the work of Claude Shannon, and the profound implications of information theory for understanding communication, complexity, and the limits of computation. I find that Shannon's insights continue to resonate across disciplines, from machine learning to software architecture, and provide a powerful lens for thinking about the flow of information in complex systems.
I also have real hobbies.