I was leading both the architecting and implementation of a document parsingpipeline and a corresponding"AI document chat". My role involved integratingtechnologies such as Qdrant vector store, Celery, and FastAPI, and makingcontributions to the deepdoctection framework, enabling efficient processing andembedding of even hard to parse documents like scanned pdfs with structured datalike tables inside. At the same time I was part of another architecture team where wehad a consulting role for a customer helping them design their SaaS architecture fromground up. This involved doing research on topics like multi cluster kubernetes andconstructing detailed ADRs.

I led the practical part of the course and prepared and presented material for students in English and German. During the course, students had to work together on several Java projects and exercises I prepared to reinforce their knowledge in the field further.

Developing FireStart Cloud.A Cloud SaaS using Angular, Typescript in the frontend and C# in a .NET microservice backend. Working within a highly driven team with a major focus on continuous improvement.