Project Partner: Prophet AI, Spring 2026
Project Summary: Modern livestock farming faces a major challenge: it is difficult to monitor the health and behavior of individual animals at scale. Subtle signs of illness, like changes in movement or eating habits, often go unnoticed, leading to delayed treatment, reduced animal welfare, and inefficient resource use. This has environmental consequences, as mortality in flocks raises the risk of infection, and increases costs to farmers due to removal and wasted feed cost.
Eric partnered with Prophet AI on their computer vision models, helping design data driven tools that analyze animal movement using computer vision. From large-scale video data he was able to extract behavioral signals, such as walking patterns and activity levels, to identify bird health. This took experimentation of methods for calculating a waddle and how to aggregate and analyze signals across time.
This work is one step in a process of transforming raw video data into actionable insights for farmers, enabling earlier detection of health issues and more targeted interventions. Improving animal welfare and reducing inefficiencies contribute to more sustainable and responsible food production.
This experience strengthened Eric’s interest in building AI-powered products that drive real-world impact, and he plans to continue developing data-driven tools at the intersection of technology, sustainability, and human-centered design.
Project Deliverables:
- Developed a data pipeline to aggregate bird movement data and compute behavioral metrics, including waddle detection signals.
- Research summary and exploratory code to identify waddle signals.
Watch Eric’s URBAN Internship lightning talk here.