At Amazon Fulfillment Technologies & Robotics, I led development and product ownership of internal robotics data-visualization applications, adding features while reducing latency several-fold and raising availability to over 99%.
I also developed Python analysis tools and Bash scripts used in robotics development, and built video-processing pipelines that generated hundreds of videos daily for manual annotation. I integrated LLM functionality into an internal tool via Amazon Bedrock.
At Forecasty.ai, a BASF-funded start-up, I developed forecasting algorithms for sales demand and commodity markets. The production models covered around 100 commodities, and I built MLOps processes for experiment tracking, model registration, containerization, and monitoring.
My earlier research at University Medical Center Göttingen involved neuropsychiatric study data and classification models; at Central Institute of Mental Health, I developed neural-network simulation models and earned a PhD from Universität Heidelberg. More recently, I’ve been training in generative AI and building personal software projects, including an iOS screen-time management app.
