PhD Scholar: ML-Driven 3D XRD Reconstruction

Lightsources.org; Grenoble; Hybride; EUR 28 000 - 38 000; Plein temps

Avantages offerts par ce poste

Relocation support

Competitive compensation

Open-source software release

Résumé du poste

The ESRF invites applications for a CFR502-PhD Student position in Grenoble to develop neural-network-driven reconstruction tools for 3D X-ray diffraction. Based at beamlines ID11, ID03, you will generate synthetic training data and train networks to accelerate reconstruction for industrial users and contribute to open-source releases.

The PhD is hosted by Université Grenoble Alpes within the Physics doctoral school.

Qualifications

• Degree (MSc, Master) in Physics, Materials Science, Engineering, or related field that qualifies for PhD enrollment in Physics at UGA.
• Strong interest in X-ray diffraction and materials characterization techniques.
• Solid background in machine learning techniques.
• Experience with Python.
• Proficiency in English (working language at ESRF).

Responsabilités

• Develop a phantom microstructure generation pipeline and validate synthetic diffraction output against real ID11 datasets.
• Train and validate neural network indexing models, progressing from box-beam to full scanning geometry.
• Benchmark automated pipeline against conventional reconstruction across materials including deformed and additively manufactured samples.
• 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 microstructure imaging community.

Connaissances

Machine learning

English proficiency

Independent collaborator mindset

Formation

MSc in Physics or Materials Science

Outils

Python

Description du poste

The ESRF invites applications for a CFR502-PhD Student position in Grenoble to develop neural-network-driven reconstruction tools for 3D X-ray diffraction. Based at beamlines ID11, ID03, you will generate synthetic training data and train networks to accelerate reconstruction for industrial users and contribute to open-source releases.

The PhD is hosted by Université Grenoble Alpes within the Physics doctoral school.