About this Technical Webinar:
FoSMo: Collaborative AI for Smarter, Shared Network Asset Management
In this session, we explore how advanced computer vision and AI techniques are transforming the monitoring, inspection, and maintenance of electricity network assets across the UK. Drawing on previous work in asset inventory, corrosion detection, and condition assessment, we show how similar approaches for different Transmission and Distribution Operators revealed a shared opportunity: many network assets fail in similar ways, highlighting the potential for a collaborative, shared-model approach.
This insight led to the creation of FoSMo (Foundational Shared Models), an industry-led project that combines image and operational data from multiple operators to train AI models that are more accurate, reliable, and actionable. The models can detect components, identify defects, and highlight corrosion, working closely with operators to ensure outputs meet their operational needs. FoSMo also addresses challenges such as secure data sharing, anonymisation, and compatibility with diverse asset management systems. A demonstration platform is already in place, allowing operators to trial models, integrate them with existing GIS and asset systems, and iteratively improve performance.
Presenter: Petar Gyurov
Petar Gyurov is Technical Lead at Keen AI, specialising in production-ready AI systems. He bridges the gap between user-facing platforms and computer vision pipelines, with over 10 years of experience in backend engineering, cloud infrastructure, and computer vision.
He holds a first-class degree in Theoretical Physics from Royal Holloway, University of London.
