PhD Position in Automated Gamma Spectrum Analysis via Hybrid Machine Learning

This doctoral position at Université Paris-Saclay, in collaboration with CEA/List and CEA/DEDIP, focuses on advancing automated gamma spectrum analysis for nuclear facility decommissioning. The research extends a newly developed hybrid machine learning and statistical unmixing algorithm to characterize concrete surfaces and classify waste contamination levels. The selected candidate will conduct feasibility studies, model model deviations, and optimize decision-making robustness for low-statistics scintillator detector measurements. The project offers a structured three-year research experience within a leading national laboratory setting.

• Advance the hybrid machine learning and statistical algorithm for concrete surface radiological characterization
• Perform feasibility studies on modeling deviations of learned models to improve decision-making robustness
• Analyze low-statistics gamma spectra acquired from scintillator detectors like NaI(Tl) and plastics
• Collaborate with research teams across CEA/List and CEA/DEDIP to integrate tools into decommissioning workflows
• Validate algorithm performance on radionuclide identification and quantification under significant environmental deformations