Research Assistant in Agentic AI Infrastructure Management

Research Assistant in Agentic AI for Infrastructure Management

This position is part of the project Neuro-Symbolic AI for Infrastructure Management, which aims to develop explainable, adaptive, and trustworthy AI systems for infrastructure management under uncertainty. Infrastructure managers increasingly need to make reliable decisions in the context of climate change, ageing assets, fragmented data, and limited monitoring information. Conventional AI methods often depend on large datasets and offer limited explainability, making them difficult to apply in safety‑critical infrastructure decisions.

This project addresses that gap by combining machine learning, structured domain knowledge, knowledge graphs, simulations, multi‑agent systems, and reinforcement learning.

The (proposed) research assistant will support the development and testing of AI agents that can codify infrastructure management knowledge from standards, manuals, inspection reports, and other technical sources into machine‑readable knowledge structures. They will also help develop and evaluate agentic AI systems that use this knowledge to reason, coordinate with simulations, and support decisions related to maintenance, inspection, resilience, and planning.

The role will involve working with machine learning methods, knowledge graphs, large language or concept models, multi‑agent systems, and decision‑support workflows.

The successful candidate will work within the Integral Design and Management Section of the 3MD Department, Faculty of Civil Engineering and Geosciences, TU Delft, and will be associated with the Digi Construct Lab. The position is suited to someone with a strong interest in machine learning and infrastructure management, and offers opportunities to contribute to prototypes, case studies, publications, and scientific papers.

• A completed or ongoing MSc degree in a relevant field such as Computer Science, Data Science, Artificial Intelligence, Civil Engineering, Construction Management, Infrastructure Management, or a closely related discipline.
• A strong interest in working at the intersection of artificial intelligence and the built environment, particularly in infrastructure and asset management, or digital twins.
• A suitable profile may include either a computer science, data science, or AI background with an interest in infrastructure systems, or a civil engineering, architecture, or construction management background with a strong interest in AI and computational methods.
• Demonstrated experience with, or clear motivation to work on, one or more of the following areas: agentic AI systems, knowledge graphs, large language or concept models, machine learning, infrastructure management datasets, simulation‑based decision support, multi‑agent systems, or reinforcement learning.
• Good programming skills, preferably in Python, and familiarity with tools or libraries used for machine learning, data analysis, semantic modelling, knowledge graphs, or AI system development.
• Experience with data processing, data structuring, or working with heterogeneous technical datasets such as standards, manuals, inspection reports, infrastructure records, or simulation outputs is desirable.
• Knowledge of infrastructure management, construction management, asset management, or decision‑making processes in the built environment is desirable, but not essential.
• Awareness of FAIR research principles, including the ability or willingness to work with research data, code, models, and documentation in a way that supports findability, accessibility, interoperability, and reusability.
• A willingness to learn new algorithms, tools, and developments in AI, including neuro‑symbolic AI, knowledge graph‑based reasoning, agentic AI, and reinforcement learning.
• A critical, creative, and proactive approach to research, with good analytical and problem‑solving skills.
• Good written and verbal communication skills in English, with the ability to document technical work clearly and contribute to academic outputs.
• Demonstrated experience in academic writing, technical reporting, literature review, or preparation of research papers is an advantage.
• Abili…