I've built automated data pipelines, analytics infrastructure, and production AI systems across Straiv, Universität Tübingen, and Shokz Technology. At Straiv, I process unstructured text through ingestion, parsing, embeddings, and Qdrant indexing, while supporting a live customer-support application that reduced average response time by 50%.
Previously, I automated Python and SQL ETL pipelines for sales, cost, operational, transaction, and review data at Shokz Technology, replacing manual reporting and building KPI dashboards. I also developed feature-engineering pipelines and dashboard indicators for a university dropout-risk model combining Bayesian Structural Equation Modelling and LLM APIs.
