Dr Mohammed Hasanuzzaman (Queen's University Belfast), Prof Christos Tachtatzis (University of Strathclyde), Prof Ilias Kyriazakis (Queen's University Belfast), Terry Canning (CattleEye), Adam Askew (CattleEye).

CattleEye has developed a “First of its kind” AI based tool which uses Machine Vision powered by Deep Learning Neural Networks to monitor dairy cattle lameness and body condition score. CattleEye is generating daily mobility and body condition scores. This technology has been academically validated and commercially adopted by almost 100 dairy farms covering 180K cows across the world in countries such as UK, USA, Germany and Argentina. The company has recently been acquired by GEA, one of the largest dairy milking parlour companies in the world and is now planning to scale its technology into a vast market of approximately 200m dairy cows worldwide.

GEA offers advanced robotic milking parlours designed to enhance efficiency, milk quality, and animal welfare on dairy farms. Their systems cater to various herd sizes and operational needs, featuring both rotary and box-based configuration. The DairyProQ is a fully automated rotary milking system tailored for large-scale dairy operations. It supports herds ranging from 600 to over 5,000 cows, with capacities of 120 to 400 cows per hour. This system employs the "In-Liner Everything" principle, performing all milking steps—stimulation, cleaning, pre-milking, milking, and post- dipping—within the teat cup, ensuring optimal hygiene and udder health. GEA's DairyRobot series offers versatile box-based milking solutions suitable for various herd sizes and barn configurations. These systems are designed for both voluntary and batch milking, providing flexibility to match farm management practice.

This PhD project will involve applying multimodal machine learning approaches in order to make better use of the generated granular data. Specifically, you will fuse CattleEye's unique multimodal data (analysing gait, behaviour, and body condition) with other relevant datasets from GEA's systems. These advanced ML approaches will allow the dairy sector to improve its ability to flag cows that require attention earlier and make more efficient use of farm staff and foot trimmers time. These models will also enable predictive diagnosis of animals with conditions such as Mastitis, Acidosis, Ketosis and Metritis.

Further research will explore how LLM models might effectively combine the existing research and best practices across the dairy industry and Artificial Intelligence research to create a new generation of explainable predictive models diagnosing conditions. This means you won't just be building a "black box" model; you'll be creating AI agents that can explain their reasoning to a farmer in plain language.

This PhD studentship offers an invaluable opportunity for the successful student to work at Queen’s University, Belfast with access to support, guidance and industry experience with local AgriFood Tech Company successful and gathering data from countries across the globe. You will gain a unique skillset at the intersection of multimodal learning, time-series analysis, and large language models, all applied to a globally significant sustainability challenge.