Mannheim, Baden-Württemberg, Germany
Currently serving as Senior AI Engineer, Image Intelligence in Danaher Corporation's Enterprise AI department—and loving every moment of it! I'm combining my passion for biology and microscopy with advanced AI techniques to develop next-generation imaging intelligence solutions. The learning curve is steep, the challenges are fascinating, and the potential for impact keeps me motivated. Being part of a team that's applying AI to advance human health feels like the perfect intersection of my interests and skills.
Led a software R&D team for Leica AIVIA, an advanced image analysis platform with GPU-accelerated n-dimensional visualization, particle tracking, neuron tracing, cell analysis, and deep learning for denoising and image segmentation. Introduced MLOps techniques focused on experiment management using MLflow, enhancing our workflow efficiency. Drove the adaptation of community deep learning models from bioimage.io for inference within AIVIA to expand the range of image analysis use cases. Additionally, led the extension of AIVIA's data exploration and visualization tools, driving greater insights and usability.
Led an R&D team focused on image processing and deep learning. We developed a modular AI deployment platform in CUDA for native image analysis, scaling across Windows workstations, Linux servers, and GPU-accelerated edge devices. Co-invented Leica's AI-powered image enhancement software for content-aware frame interpolation, pushing the boundaries of high-speed live cell recording [https://tinyurl.com/dse-mara]. Designed and executed AI-driven technology roadmaps for advancements in image enhancement, information extraction, automation, and knowledge discovery, laying the groundwork for future innovations and competitive advantage.
Global product manager of numerous products comprising mainly confocal microscopy. I introduced a game changing photon counting hybrid detector called "HyD" to the TCS SP5 and launched the SP8 confocal platform. It ran for 8 years very successfully. I also collaborated with external partners to create a CUDA-enabled deconvolution solution called "HyVolution" based on HyD. This paved the way to what later became Leica's trademarked Lightning and Thunder products by demonstrating the potential of software-enhanced detection systems to push the boundary for gentle imaging of live specimen. Towards the end of my tenure in this role I built Leica's first POCs for AI-powered microscopy.