PhD in Machine Learning and 3D Tomography for Turbine Blade Inspection

A fully funded CIFRE PhD position is available at CentraleSupélec, Université Paris-Saclay, in collaboration with Safran and hosted at the MICS lab. The research focuses on developing advanced machine learning methods for the accelerated 3D analysis of tomographic measurements of next-generation turbine blades at various manufacturing stages. The project aims to automate the inspection process of high-pressure turbine blades, which are critical components in aeronautical engineering, by leveraging signal processing, diffusion models, visual transformers, and 3D discriminative models. The research will address the detection of metallurgical anomalies such as porosity, shrinkage cavities, and cracks, with the goal of meeting stringent aeronautical quality control standards. The PhD will be co-supervised by Myriam TAMI (Associate Professor, CentraleSupélec), Clément Remacha, and Julian Betancur. The project will involve evaluating and improving existing models, proposing new strategies for 3D tomographic inspection, and integrating these methods into a complete industrial processing chain. Candidates will investigate parallelization techniques and model optimization strategies, including teacher-student knowledge distillation. The research will culminate in a rigorous evaluation protocol using quantitative indicators like F1 score and ROC curves, aiming to achieve high detection probability and minimize false negatives. This opportunity is ideal for candidates with a strong background in machine learning, signal processing, or related fields, and an interest in industrial and aeronautical applications. The position is fully funded through the CIFRE program in partnership with Safran. The recruitment process is ongoing, and interested candidates are encouraged to apply as soon as possible.