01 · Profile
Design thinking, engineering discipline
- Location
- Cologne, Germany
- Experience
- 25+ years
- Current
- Science Media Center Germany
My background combines design thinking with software engineering, covering software development, data
engineering, semantic technologies, knowledge graphs, AI and machine learning.
I work on complex technical problems where different disciplines need to come together: software
architecture, data and knowledge modelling, AI/ML, information architecture and user-facing applications. My
goal in this work has always been to make complex information and capabilities easier to understand, use and
work with.
02 · Focus
Current areas of focus
- 01Generative AI
and LLM-based systems, including RAG, AI agents, prompt and workflow design
- 02Knowledge engineering,
semantic technologies and knowledge graphs
- 03Search and retrieval
across structured and unstructured data
- 04Database systems and
data architecture
- 05Software architecture
for complex applications
- 06AI-powered tools and
interfaces
- 07Product design, UI/UX
and information architecture
03 · Trajectory
From design to AI engineering
Twenty-five years, four disciplines, one thread: making complex information usable.
-
-
Computer Game Designer
calculus Softwareengineering · Freelance
Cologne · 1996–1998
Game design and development at a small Cologne studio.
-
Designer
Freelance
Cologne/Bonn · 1998–2004
Comprehensive design services across print and digital media.
- Corporate identities and large-scale event conceptualization
- 3D modeling, visualization and exhibit design
- Multimedia applications, video editing and early web development
Origins 1995 — industrial design at U. Wuppertal · exhibit-design internship at
Tetrad Design, Annapolis, MD.
-
-
Research Associate
U. Cologne · Institute of Digital Humanities (HKI) · Full-time
Cologne · 2007–2009
Digital-library infrastructure for manuscript and copyright-history collections
— ENRICH and “Primary Sources on Copyright History”.
- Browser-based content management for contributing authors; backend XML database implementation
- Schema conversion to Dublin Core and TEI P5 via XSLT; OAI-PMH interfaces
- Frontend concept and design
-
Lecturer
U. Cologne · Part-time
Cologne · 2007–2010
Practical exercises in C++; introductory programming courses in Processing/Java;
faculty tutorials.
-
Research Associate
U. Cologne · CoDArchLab, Institute of Archaeology · Full-time
Cologne · 2009–2011
Core development on the Arachne archaeological database, incl. the Hellespont
project integrating Arachne and Perseus.
- LAMP-based core system contributions
- Standard schemas and ontologies — Dublin Core, CIDOC CRM, METS/MODS — for data transfer and
aggregation
- Prototyped applications on those interfaces
-
Software Developer
Cambridge Faculty of Law · Freelance
Cambridge, UK · 2011–2013
Rebuilt the digital archive “Primary Sources on Copyright History (1450–1900)” —
see Selected work.
-
-
Software Developer
HRG · Freelance
Cologne · 2012–2015
REST interfaces for bidirectional data transfer between travel booking systems,
plus import-validation monitoring with a browser-based interface.
-
Software Developer
Heimkomfort · Freelance
Cologne · 2013–2015
Complete retail infrastructure for an online commerce business on Magento —
owned end to end, platform to DevOps.
- Self-managed open-source web stack, from setup through production operation
- Custom platform modules and browser-based back-office tools
- Database administration, DevOps and continuous maintenance
-
Software Engineering | Data Engineering
fedger.io · Full-time
Cologne · 2016–2019
Data-driven products for the hospitality industry.
- End-to-end ETL/ELT pipelines across relational and non-relational stores
- Multilingual knowledge graph integrating public linked data with proprietary industry data
- NLP/ML extracting structure from domain-specific text; data fusion across text, images,
point-of-sale, geolocation and social media
- Interactive web visualizations for non-technical stakeholders
-
04 · Selected work
Selected work
Three projects that show the thread: structure the information, then make it usable.
Bridging the Communication Gap
2024–present · Science Media Center Germany × U. Freiburg
AI-supported science communication with the Faculty of Environment and Natural
Resources (University of Freiburg), serving forestry, environmental and sustainability sciences.
- Hybrid semantic search over vector-database infrastructure — embeddings, lexical scoring and neural
reranking across data sources
- LLM-powered enrichment following RAG methodology; multi-agent architectures with tool servers on the
Model Context Protocol
- TypeScript/React frontends for conversational AI, faceted search and monitoring, on token-authenticated
REST APIs with SSE streaming
suite.bricomgap.uni-freiburg.de →
Augmented Science Journalism
2021–2023 · Science Media Center Germany
Knowledge-graph and AI solutions for editorial content analysis and processing —
backend infrastructure, data platforms and APIs.
- Knowledge graphs grounded in semantic-web standards, integrating editorial content with public
linked-data repositories and ontologies
- LLM infrastructure operated end to end, from self-hosted model serving to orchestration and low-code
workflow automation
- Polyglot data infrastructure — relational databases, GraphQL interfaces and vector stores — for hybrid
search over structured and unstructured content
Primary Sources on Copyright History, 1450–1900
2007–2009 · 2011–2013 · U. Cologne × Cambridge Faculty of Law
Digital archive of historical copyright documents — conceptualized in Cologne,
rebuilt in Cambridge six years later. The archive is still online.
- Rebuilt system architecture on CouchDB and RESTful principles, with an XML→JSON migration pipeline for
legacy data
- Full-text search with Apache SOLR; OAI-PMH metadata interface on Dublin Core standards
- Browser-based CMS for research contributors; integrated external sources (Wikipedia, Google Books); URL
persistence and permalinks
copyrighthistory.org →
