COME, working on the edge between computer science and animal science, develops theoretical machine learning algorithms alongside practical, data-driven solutions for challenges in agriculture and veterinary medicine. To this end, we evaluate and develop a range of modern sensor technologies and data analysis methods. Recurring themes are applicability and cost-efficiency. We aim to ensure that tools can be used by non-experts at reasonable cost, even though complex technology runs in the backend. Achieving this requires interdisciplinary collaboration: end-users need to understand the technical requirements, and developers need to understand the needs of the end-users. In this way, COME developed a range of applications, from assisted telemedicine and teletriage to the automatic analysis of images, video, and sensor data. Concrete examples include drone-based phenotyping for plant breeding, computer-vision measurement of livestock, smartphone-based diagnostic tools for pet owners, and route-planning systems that help locate the nearest available veterinary care. Across these efforts, the goal was the same: optimizing resources, improving animal welfare and agricultural sustainability, and providing the basis for further innovation. Bridging methodological research and real-world deployment, COME remains a key player in computational veterinary medicine, while also working to bring these advances into modern teaching.