Selected work

Assessed the resilience of a U.S. road network to combined flood and earthquake hazards, estimating direct and indirect losses using hazard maps and vulnerability models, and linking recovery times to traffic performance to support national resilience standards.

• Built flood & earthquake hazard impact simulation in Python.
• Linked road recovery times to traffic disruption metrics.
• Evaluated long-term risk and resilience policy options.

A decision-support tool for transport risk management in South Asia, performing probabilistic risk assessments for natural hazards to guide infrastructure investment and emergency planning.

• Developed a modular Python tool for hazard and traffic flow simulation.
• Modelled flood/earthquake risk across South Asian networks.
• Supported resilience planning and investment prioritisation.

A system to improve travel for mobility-impaired passengers at Euston Station, using Digital Twin technology to provide real-time updates, personalised routing and enhanced disruption support.

• Developed real-time speech-to-text and translation for live announcements.
• Integrated personalised routing with the Digital Twin platform.
• Enabled inclusive support for mobility-impaired passengers.

Pedestrian flow modelling for the repositioning of Customer Information Screens at Euston Station, ensuring the new layout improved flow and reduced congestion.

• Performed pedestrian flow analysis for screen placement.
• Analysed LiDAR data to identify low-disruption locations.
• Used simulation to assess the impact of signage changes.

Sourced and benchmarked points-of-interest (POI) data for restaurants, cafés, hotels and laundries across Attica, Thessaloniki and Crete, to support a field-sales and site-selection campaign for a WEST client.

• Benchmarked Google Places against HERE on coverage, accuracy and cost.
• Delivered categorised, geolocated datasets ready for mapping and field targeting.
• Scoped and priced tens of thousands of POIs across three regions to a tight deadline.

An AI-based decision-support tool for real-time disruption recovery at Euston Station. Using over 100 million data points, it generates adaptive recovery strategies aligned with DfT and Network Rail priorities.

• Built the core AI optimisation model for disruption recovery.
• Combined meta-model scheduling with pedestrian microsimulation.
• Achieved a 20% flow conflict reduction in station simulations.

A proof-of-concept Station Living-Lab at Euston with a real-time Digital Twin to support HS2, WCP and Network Rail, informing station design, operations and asset management.

• Led modelling for the real-time Digital Twin PoC at Euston.
• Used AI-driven analytics to monitor and forecast passenger flow.
• Provided insights for HS2 design and operational decisions.

Explored the technical feasibility of a Digital Twin for real-time estimation and simulation of end-to-end passenger movements, predicting station occupancy and boarding/alighting figures.

• Led transport modelling and data fusion for King's Cross–St Pancras platform.
• Built and calibrated pedestrian simulations using observed and forecast data.
• Delivered insights for station flow management and customer experience.

A feasibility study for a national-scale Digital Twin for UK rail, developing real-time monitoring, passenger flow prediction and what-if simulation tools to support operational resilience.

• Led modelling and data analysis for a national rail Digital Twin feasibility study.
• Developed real-time monitoring and short-term prediction methods for station flows.
• Built simulations to support demand estimation and disruption alerts.

A prototype open-architecture system for urban transport and environmental management, integrating traffic and air pollution monitoring, piloted in Nicosia, Chania and Kos.

• Developed a reinforcement learning module for traffic signal optimisation.
• Integrated real-time traffic and pollution data.
• Deployed and tested a large-scale system in Nicosia.