Marat Dusaev, a graduate of the Institute of Economics and Management of Tomsk State University, together with his team, developed and implemented an artificial intelligence-based system for digital monitoring of the health of farm animals on the university’s supercomputer. The project received the IV International University Prize in Artificial Intelligence and Big Data – Gravity.
Gravity is an international competition for university IT projects. Its organizers are the Presidential Academy, Tomsk State University and the University Consortium of Big Data Researchers. The competition is held annually. In 2026, the organizing committee received 243 applications – 20% more compared to 2025.
The VETAI project of Marat Dusaev and his team won in the “Promising Project” category in the nomination “High Social Effect”, which marks practice-oriented projects that use AI or big data for social purposes.
The VETAI system is designed for dairy farms to monitor the health of farm animals via video stream. It uses computer vision and neural network models to detect early signs of disease and deviations in livestock behavior 24/7 without the use of sensors or boluses: signs of lameness, mastitis, bedsores and stress.
— On large farms, a veterinary specialist cannot constantly monitor every animal. However, early signs often appear before the disease becomes apparent. The AI system translates these weak signals into digital events: the animal moves less, lies differently, changes activity, reacts differently to the herd or feeding, – says Marat Dusaev.
Video analytics, herd control, animal events, feeding, management dashboards and predictive analytics are combined in one interface for the manager, veterinarian and farm team.
The key factor in the development of the VETAI system was the infrastructure of the Institute for Big Data Analysis and Artificial Intelligence of TSU. In particular, the project team performed resource-intensive tasks on the TSU supercomputer. The startup has a high-performance node of the supercomputer center for training neural network models and building complex AI-pipelines.
— Modern artificial intelligence projects cannot be developed without a powerful computing infrastructure. Training neural network models, processing large amounts of data and working with video streams require huge resources. TSU supercomputer center provides researchers and technology teams with access to such resources, and thereby helps to move faster from a scientific idea to a ready-made solution, – commented Vyacheslav Goiko, director of the Institute for Big Data Analysis and AI.
According to Marat Dusaev, the VETAI team is currently preparing to test the system in real farms. Thus, the project team has entered into an agreement with Gatchinskoye JSC in Leningrad Region and is already preparing the system for launch on this farm.