PhD Position Statistical and Machine Learning Descriptors for Metallic Microstructures

About the position

Job description

Join us to explore the complex nature of microstructures through a unique intersection of applied mathematics, machine learning, and materials science

Job description

The properties of metals are shaped by their microstructures: intricate three-dimensional arrangements of grains, phases, and interfaces that emerge during processing. Understanding this internal architecture is key to designing better materials, optimising manufacturing processes, and developing predictive digital twins. Yet, despite major advances in materials characterisation and modelling, we still lack ways to capture the full complexity of microstructures in representations that are both mathematically rigorous and physically interpretable.
Can we develop new mathematical representations of metallic microstructures that capture their complexity while revealing their connection to material properties?
This is the central challenge of this PhD project. Working at the intersection of materials science, applied mathematics, and machine learning, you will explore new ways to describe, compare, and understand complex metallic microstructures. The ambition is to move beyond conventional descriptors, such as average grain size or phase fraction, while retaining the physical interpretability that is often missing from purely data-driven approaches.
You will investigate how the geometry, spatial organisation, topology, and crystallographic characteristics of microstructures can be captured in compact, information-rich representations. By combining mathematical insight with modern machine learning, the project aims to uncover the underlying organisation of microstructures: what makes two microstructures similar or different, which structural characteristics distinguish them, and how these characteristics relate to material properties.
A particularly exciting aspect of this project is the opportunity to bring together ideas and approaches from traditionally distinct disciplines. From statistical analysis, topological data analysis, and manifold learning to advanced microstructure characterisation and computational materials modelling, you will explore how different mathematical and physical perspectives can complement each other. Rather than simply applying existing techniques, you will have the opportunity to develop new connections between these fields, with the aim of establishing a more comprehensive and physically meaningful framework for understanding microstructures.
The project will focus primarily on steel microstructures, contributing to the development of more sustainable and efficient steel production. It will be embedded in a larger research consortium dedicated to green steel, providing opportunities to connect fundamental research with real-world challenges in materials production and processing. At the same time, the mathematical concepts developed in this project will have potential applications well beyond steel, across a broad range of polycrystalline metallic materials.
You will conduct your research at TU Delft, jointly embedded in the Department of Materials Science and Engineering and the Statistics Group of the Delft Institute of Applied Mathematics. This interdisciplinary environment offers a unique opportunity to develop fundamental mathematical and computational ideas while addressing challenging questions in physical metallurgy. You will benefit from expertise across both disciplines and opportunities to engage with academic and industrial partners throughout the metallic materials value chain.

What we offer:

A 4-year fully funded PhD position at TU Delft.
Collaboration with international experts from diverse disciplines.
Access to cutting-edge computational and experimental facilities.
Supervision of MSc researchers.
Mentorship and career development support.
The position is jointly embedded in the research groups of:
Dr. Martina Vittorietti
Dr. Maria Santofimia
Dr. Sid Kumar
Job requirements
A Master’s degree in applied mathematics, statistics, materials science, applied physics, mechanical, aerospace, or civil engineering, or a related field.
Experience in statistical methods, data analysis, and/or machine learning for scientific applications is preferred.
Strong programming skills (e.g., Python, PyTorch, TensorFlow, or R).
Interest in experimental materials characterisation, including techniques such as microscopy.
Curiosity and willingness to explore beyond your own discipline. If your background is in mathematics or statistics, you should have an interest in learning about materials science. Conversely, if you come from a materials science or engineering background, you should be eager to deepen your understanding of applied mathematics and statistical methods. You are not expected to be an expert in both fields from the outset.
Motivation to tackle challenging interdisciplinary research questions and connect ideas across disciplines.
Strong communication and teamwork skills, with an enthusiasm for collaborating with researchers from different scientific backgrounds.
TU Delft (Delft University of Technology)

This listing was collected from a public source and is reproduced here for information only. Always confirm the details on the original posting before applying.
View the original posting

Similar positions