PhD Student in Neural Networks for X-ray Diffraction

The European Synchrotron Radiation Facility (ESRF) in Grenoble, France, invites applications for a fully funded PhD position focused on creating neural-network-driven reconstruction tools for Scanning 3D X-ray Diffraction (S3DXRD). Operating at the intersection of synchrotron science, materials characterization, and machine learning, the successful candidate will generate synthetic diffraction data from simulated microstructures to train models that replace expert-intensive analysis methods. The project involves methods development, experimental validation, and open-source software release in collaboration with beamline scientists and the Algorithms & Scientific Data Analysis group. The doctoral degree will be hosted by Université Grenoble Alpes within its Physics doctoral school.

• Develop a phantom microstructure generation pipeline and validate synthetic diffraction outputs against real experimental datasets.
• Train and validate neural network indexing models progressing from box-beam to full scanning geometries.
• Benchmark the automated reconstruction pipeline against conventional methods across diverse deformed and additively manufactured materials.
• Collaborate with crystal plasticity simulation groups to ensure training microstructures capture realistic orientation gradient fields.
• Package, document, and release an open-source reconstruction toolkit for the broader diffraction imaging community.