Lead the integration of LLMs and generative AI with Retrieval-Augmented Generation (RAG) technology for R&D reports, significantly improving the accuracy and efficiency of information synthesis and knowledge discovery in research processes.Pioneered the application of active learning, accelerating the development of new materials by optimizing experimental designs and reducing time-to-discovery.Computing predictive models and extracting insights with ML techniques from float glass process and numerical simulation data. CI/CD with Gitlab and Kubernetess.Supporting the digital transformation of the research center by implementing new databases and developing dashboards for experimentalists.Developing a home-made Bayesian fitting procedure to estimate glass stress resistance with Weibull distributions.Providing support to junior data scientists and advising the management to develop data-driven approaches.Developing customized real-time visualization tools for float glass lines to increase reactivity reducing lost in production.

Virtual screening of new organic materials (Clean Energy Project DB) with high-throughput ab initio computational methods (DFT-MBPT) for green energy applications (Li-ion batteries and organic solar cells). Data mining of a large database of materials properties (Materialsproject DB) with deap learning models and graph networks. Development of Python libraries in the framework of the Materials Project to perform high-throughput simulations (github.com/materialsproject)Data mining of a large database of materials properties with deap learning models and graph networks (use of API and MongoDB)Machine learning project for an industrial client: data preparation and cleaning, EDA and data modeling (linear regression, Ridge, lasso, elastic net, PLS, , gradient boosting and random forest regressors, neural network, ...) for solid state material properties.