PhD Studentship in Computer Science: Autonomic Computing and Agentic AI for the Internet of Things and Edge Systems

Award Summary

100% fees covered, and a minimum tax-free annual living allowance of £20,780 (2025/26 UKRI rate). Additional project costs will also be provided.

Overview

The Internet of Things (IoT) and Edge Computing are undergoing a major transformation. Systems that once relied heavily on cloud-based processing and passive data collection are evolving into distributed networks of intelligent, decision-making devices. Operating at the edge imposes strict constraints, including limited compute and memory resources, intermittent connectivity, strong privacy requirements, and tight energy budgets.

Under these conditions, autonomic computing becomes essential. Devices must be capable of self-configuration, continuous optimisation, fault recovery, and adaptation to changing environments without external control. Autonomous agents that can perceive, reason, plan, act, and learn, together with self-configuring, self-healing, and self-optimising behaviours, provide the foundational principles for this shift.

Recent advances in small language models, TinyML frameworks, sparse neural networks, and microcontroller-grade accelerators now make it feasible to deploy sophisticated reasoning and self-management directly on constrained devices alongside their primary applications.

Research Scope

This research focuses on the convergence of Agentic AI and Autonomic Computing at the extreme edge, investigating how to design, deploy, and manage IoT systems that are genuinely self-managed and resilient. The goal is to enable autonomous, reliable operation across diverse domains, including smart buildings and cities, healthcare devices, industrial IoT, and connected or autonomous vehicles.

Classical AI and distributed systems solutions cannot be directly applied at the edge due to severe resource limitations. A key challenge lies in enabling agents to continuously adapt their knowledge and behaviour in dynamic environments while keeping decision-making processes explainable. This motivates the use of lightweight, logic-based machine learning approaches. In addition, agents must support collective decision-making to achieve system-wide optimisation rather than isolated, local improvements.

Finally, collaboration among edge agents enables decentralised decision-making and global optimisation without continuous cloud dependence. By sharing local observations, learned models, or inferred knowledge, agents can collectively detect anomalies, coordinate actions, and adapt more effectively. Such collaboration must remain lightweight, scalable, privacy-aware, and resilient to intermittent connectivity.

Aims & Objectives

The research aims to investigate the convergence of Autonomic Computing and Agentic AI for self-managed IoT/Edge systems.

Supervision Environment

The candidate will be trained and work with experts from the Intelligent Systems Research group (with extensive experience in deploying machine learning models for edge devices and microcontrollers and building simulation environments for IoT), and our extensive network of international partners. Research will be done in the cooperation with Literal Labs, a NU spin-off from the School of Engineering that is working on logic-based ML models and with National Edge AI Hub.

Number Of Awards

1

Start Date

1 October 2026

Award Duration

4 years

Application Closing Date

15 February 2026

Sponsor

EPSRC

Supervisors

Dr Tomasz Szydlo , Dr Devki Nandan Jha , Dr Rishad Shafik

Eligibility & How to Apply 

For eligibility criteria and how to apply please visit our website

Contact Details

Dr Tomasz Szydlo