At Vigensis, we're building VIA. an AI platform combining curated expertise with highly specialized agents, autonomous workflows and human insight. As head of engineering, I spend most of my time coordinating work across our agentic workforce while evolving the architecture of VIA. VIA brings - curated expertise in all stages of a product's life cycle, from ideation to market fit - specialized agents supporting every aspect of a product's life cycle, doing much of the heavy lifting - autonomous workflows that enable us to scale our valuable human expertise VIA is designed to - work with greenfield products as well as legacy systems - run AI agents at scale in a flexible, isolated, robust harness - support multiple LLMs from different vendors, targeting specific models for specific tasks - optimize token consumption towards achieving optimal output - facilitate sovereign deployments in regulated environments Today we use VIA for our own product ideas and first customer engagements, and of course, to evolve VIA itself. Our goal is to open it up for early adopters soon, and eventually, make it available for everyone. Redefining product development the way it should be! VIA can work with your technology stack of choice. It is built with Elixir, Python, TypeScript, Postgres and Qdrant.

The navify Algorithm Suite is a cloud-based platform bringing medical algorithms from Roche and its partners into clinical routine. It provides labs and hospitals with direct workflow integration with their EMR/LIS, facilitating reliable and more efficient processes. Platform Impact: - Serves hospitals and laboratories globally, processing patient data to facilitate faster clinical decision-making and earlier diagnosis - Delivers a comprehensive library of medical algorithms providing patient-centric insights for oncology, cardiology, infectious diseases and other conditions, supporting physicians in their daily work - Integrates Roche and partner algorithms into digital workflows and existing hospital and lab systems (EMR/LIS) Responsible for: - Architecting a scalable cloud platform serving enterprise healthcare environments globally - Leading the architecture team as the product scaled from initial deployment to a mature global platform - Ensuring overall architecture aligns with customer needs, business needs and regulatory requirements - Identifying organizational needs and recommending team structure adjustments through multiple growth stages - Guiding cross-functional work across engineering, product, business, and implementation teams - Coordinating technical work accross multiple implementation teams - Conducting privacy, security and performance assessments and ensuring that findings are addressed - Driving technical strategy and architectural runway for platform evolution and partner ecosystem in highly regulated healthcare domain Technologies: Cloud Platforms, AWS, Kubernetes, Istio, AI/ML Integration, Healthcare Interoperability (HL7/FHIR), Microservices, Enterprise System Integration

LGI-Flag is an ML algorithm that uses demographic information and lab test results to provide personal scores to indicate individuals with higher probability of harboring colorectal cancer (CRC) compared to the general population. - Re-designed the architecture for the next release of LGI-Flag to support distinct product configurations for different jurisdictions, while optimizing the technical fit with the navify Algorithm Suite. - Reworked technical documentation and detailed designs to facilitate resubmission under IVDR (now class C) that included an increase of the Software Safety Class. - Implemented the necessary changes, including an extensive baseline test set for comparison with the previous version, ensuring medical compatibility. - Designed and conducted extensive performance tests to ensure predictable throughput and latency for population screening use cases of up to 1M patients per run. Made intense use of the Virtual Architect developed earlier while fine-tuning its implementation, achieving a significant gain in efficiency, reducing development time with measurable improvements in quality and test coverage. Technologies: C#, Python, Typescript, ML/AI, AWS, Medical Device Software

Designed and implemented a Virtual Architect - an extended coding agent able to support architects and engineers in their daily activities, working across repositories for multiple products. Investigated multiple alternative approaches (e.g. LangGraph, Crew AI, Copilot, Bedrock, etc.) by implementing and comparing prototypes to determine best fit for the organization and its requirements. Then implemented a production ready agent and tested it on the navify Algorithm Suite code base (a complex globally available clinical decision support platform spanning dozens of repositories). The Virtual Architect supports the team with - Architectural reviews and compliance assessments - Impact and risk assessments for proposed changes - Elaborating and comparing alternative technical solutions - Drafting ADRs and updating architectural documentation - Writing technical requirements, specifications and test cases - Implementing changes across repositories The Virtual Architect integrates with - AHA for roadmaps and product level epic and features - Code Beamer for formal requirements, specifications and test cases - Jira for team level work package breakdowns - Confluence for architectural documentation - SonarQube for static code quality and security analysis - GitHub and GitLab for code repositories

The navify Algorithm Suite is a cloud-based platform bringing medical algorithms from Roche and its partners into clinical routine. It provides labs and hospitals with direct workflow integration with their EMR/LIS, facilitating reliable and more efficient processes. Designed the emerging architecture of navify Algorithm Suite from early concept stage onwards, leveraging existing capabilities where appropriate and identifying new reusable components needed across products. Worked closely with other architects and engineering teams to realize these components. Partnered intensively with regulatory specialists to design a compliant architecture with a patented approach for segregation and isolation of medical device functions from generic platform functions, enabling independent and efficient lifecycle management. Conducted a deep exploration with Helios hospital chain, exploring the integration of cardiac algorithms at the Leipzig Heart Center, both from a clinical workflow and a systems integration perspective. For its MVP: - Defined the scope: minimal functionality focused on what's needed to serve the first clinical use cases. Designed for security, reliability, and scalability from the start. - Designed the technical architecture: ready to start small while supporting multi-tenancy, fit for global deployments. - Coordinated the implementation accross a multidisciplinary team of ~12 people, delivering a production-ready platform in <6 months while managing significant scope changes during development. The first production deployment proved the concept, demonstrated team capability and validated market need. This success unlocked the additional investment for further platform evolution. Technologies: C# · AWS · Kubernetes · Istio · Cloud Architecture · GitOps

Worked in various roles in IT. Mostly in healthcare, but also in other sectors. See my linkedin profile for more information, feel free to reach out if you want to learn more.