Abstract: Advances in sensing, machine learning and robotics are transforming the way complex workplaces are monitored, understood and optimized. From cyber physical systems operating in emergency environments to transport hubs and complex industrial environments, the ability to accurately localise people and assets, track activities, and infer high-level semantic context is becoming a critical enabler for intelligent and adaptive workplaces.

In this talk, I will present how we have leveraged heterogeneous sensor modalities to develop robust localisation, tracking and semantic scene understanding systems, highlighting a range of diverse challenges from algorithmic and system design to edge computing and human-machine interaction.

Bio: Professor Niki Trigoni is Professor at the Oxford Department of Computer Science, heading the Cyber Physical Systems Group, and a Fellow of the Royal Academy of Engineering. Her interests lie in the tight integration of sensing and machine intelligence for context inference, control and human-machine interaction. She has applied her work to a number of application scenarios, including mobile autonomy, asset monitoring, and localisation systems for emergency situations, as well as workforce safety and efficiency.

Trigoni founded and served from 2014 to 2019 as Director of the Centre for Doctoral Training on Autonomous and Intelligent Machines and Systems. Driven by her passion for research translation, in 2015 she founded Navenio Ltd, a deep-tech Oxford spinout focused on infrastructure-free indoor positioning, recognised as a 2020 KPMG Best British Tech Pioneer. In 2020, she won the CTO of the Year award at the UK Women in IT Awards for translating positioning technologies into real-world healthcare efficiency solutions.

Abstract: Agentic AI is driving higher token volumes and compute infrastructure demands. At the same time, advances in both model efficiency and hardware mean capabilities that recently required datacenter-hosted models can increasingly run on edge devices.

This talk examines how these trends are reshaping where intelligence executes. We will discuss the benefits of intelligence distributed across datacenters, edge infrastructure, and endpoint devices, with workloads dynamically routed according to capability, latency, cost, privacy, and energy constraints - with the goal of AI systems that combine frontier-scale intelligence with rich local context, bringing together the strengths of both datacenter and edge computing.

Bio: Dominic Pajak is senior director of edge AI at Arm, where he focuses on shifting AI inference closer to data and users, enabling new classes of private, real-time, context-aware applications while reducing energy and system costs. He works with leading semiconductor partners and device vendors on solutions that balance performance, power, and cost, and has launched products now shipping in billions of units. At Arduino, he worked to accelerate on-device AI education and innovation through work with Google and Harvard University. Dominic holds a PhD in Computer Science from the University of Leeds, UK.