EPSRC FIBE3 CDT PhD studentship with Network Rail: A Generative AI Framework for Optimising Capital Allocation in Track Asset Renewal Specifications

This is a four-year (1+3 MRes/PhD) studentship funded through the Cambridge EPSRC Centre for Doctoral Training in Future Infrastructure and Built Environment: Unlocking Net Zero (FIBE3 CDT).

The project is funded in collaboration with Network Rail, the entity responsible for the operation and maintenance of the Great Britain’s railway infrastructure, with an extensive network spanning thousands of miles and its complex web of tracks, stations, signalling systems and more, and is fully committed to advancing research and innovation in the field of infrastructure and built environment to enhance efficiency, safety and sustainability.

As we navigate 2026, the global rail industry faces a systemic paradox: an unprecedented abundance of sensor data juxtaposed against rapidly aging infrastructure and increasingly constrained capital budgets. Traditional track renewal specifications remain siloed and manual, often failing to account for the non-linear impacts of climate resilience, carbon efficiency, and long-term lifecycle costs. To maintain a sustainable network, we must transcend "predictive" maintenance—which merely forecasts failure—and move toward "prescriptive" asset management.

This project aims to develop a pioneering Generative AI framework designed to autonomously synthesize and optimize capital investment strategies. By leveraging Large Language Models (LLMs) and Graph Neural Networks, the candidate will create a system capable of interpreting complex engineering standards, historical maintenance logs, and real-time telemetry to generate high-fidelity renewal specifications.

The core of the research will focus on:

•     Multi-Objective Optimization: Balancing immediate structural necessity with 30-year carbon targets and budgetary ceilings.

•     Generative Policy Design: Using AI to "draft" investment scenarios that traditionally take months of manual cross-referencing.

•     Digital Twin Integration: Linking generative outputs with live physics-based models to validate the structural viability of AI-proposed specifications.

This project offers a unique opportunity to work at the intersection of the University of Cambridge’s Department of Engineering, ensuring your research has a direct pathway to industrial adoption and policy influence.

For project-specific enquiries please e-mail Professor Mark Girolami (mag92@cam.ac.uk).For general enquiries, please email cdtcivil-courseadmin@eng.cam.ac.uk.

Applicants should have (or expect to obtain by the start date) at least a high 2.1 degree preferably at Masters level in Engineering, Data Science, or Applied Mathematics. Candidates are expected to possess strong computational skills and a desire to apply cutting-edge AI to the physical world’s most pressing "Grand Challenges".

Fully-funded studentships (fees and maintenance) are only available for eligible home students in the first instance. A limited number of international students can be considered for funding at a later stage in the recruitment process.

Further details about eligibility and funding can be found at:

https://www.ukri.org/councils/esrc/career-and-skills-development/funding-for-postgraduate-training/eligibility-for-studentship-funding/https://www.postgraduate.study.cam.ac.uk/finance/fees

https://www.cambridgetrust.org/scholarships/

Applications should be made online via the University of Cambridge Applicant Portal: https://www.postgraduate.study.cam.ac.uk/courses/directory/egegpdfib  stating project title and supervisor’s name. Please note there is a £20 application fee.

Early applications are strongly encouraged as an offer may be made before the stated deadline.

The University actively supports equality, diversity and inclusion and encourages applications from all sections of society.