Two-dimensional detectors are advanced technology recently adopted in X-ray material characterization. Optimizing their use requires software development for data management and processing.

However, their potential is limited by integration with commercial systems using closed-source software. In this project, I contributed to designing and optimizing routines for processing 2D detector data.

I addressed two key challenges: "tiling" and different integration methods required for materials to obtain an I(2θ) curve representing diffracted intensity variation.

I worked with two detectors: the Pilatus100k for polycrystalline materials and the Pixcel3D for oriented crystalline thin films.

This project was realised during the internship that concluded the final year of my Higher National Diploma in Applied Physics & Instrumentation

This project aims to enhance the detection of chemical changes during mechanochemical syntheses through the analysis of acoustic signals. The focus was on converting existing MATLAB code to Python to improve accessibility. Two main techniques were used: an energy-based method and kurtosis analysis to detect sudden changes in sound, which signal chemical transformations. The results show that these techniques are effective in identifying significant state changes in the signals, providing a solid, reusable framework for future acoustic monitoring applications.

This project was done in the context of a 2 months internship

In this project, with a team of 4, our goal was to developed a custom Python assembler, focusing on lexical and syntactical analysis of regular expressions and Python instructions. The goal was to build a tool capable of generating and executing Python bytecode through a personalized assembler. This project allowed me to delve into concepts such as compilation, assembly, code source analysis, and the functioning of a Python virtual machine (VM)

This project, developed in collaboration with fellow student Matthias Schedel, aims to create a predictive model that leverages both machine learning and generative AI (GAI) techniques. The model skillfully employs linguistic features extracted from text data to predict dementia diagnoses. The database utilized for this project is sourced from DementiaBank. Ultimately, I completed a preliminary version of the deep learning model RoBERTa, in addition to fine-tuning several machine learning models, including Random Forest, K-Nearest Neighbors and a Logistic Regressor.

This project involved creating a model to predict various parameters related to the energy consumption of the GreEn-ER building. The focus was on applying machine learning algorithms, including polynomial regression with Lasso regularization and the Random Forest Regressor, to analyze building data from 2017 to 2019.