Prof. Christian Conrad's research has always been driven by the automatic generation of cell phenotypes in microscopes. He pioneered machine-learned recognition of cell morphologies in high-throughput biological applications as an integral part of microscopic systems. In particular, he developed software support and APIs of various wide-field and confocal microscopes to trigger online time-lapse or spatial resolution based on classified cell images in perturbation screens. Recently, he began combining single-cell sequencing techniques with image-based screening, initially in organoids. With the advent of in situ sequencing directly on samples, he focused his studies on microtissues and tissue sections imaged in 3D using light sheet microscopy (SPIM). His goal is to help other international groups apply these new research methods, including deep-learning methods, and to make such technologies useful for diagnosing patients.