Pau Bujons

– At CNRS in Paris, I am the main architect, developer, and maintainer of AstroGeoFit, an open-source Python tool developed with geologists, mathematicians, and astrophysicists to derive time-depth functions from geological data. – The tool combines Genetic Algorithms and Bayesian analysis, and I built it following good OOP practices with a focus on maintainability, documentation, and usability. – It also includes preprocessing tools for large datasets, such as detrending, outlier detection, and missing data imputation with KNN, as well as visualizations built with Matplotlib. – I also developed a PyQt GUI for non-technical users, an HPC parallelization system with Dask that reduced computing costs by up to 70%, and a full CI/CD and testing environment using GitHub, GitLab, Snakemake, Docker, and Sphinx. – The tool is now used by several scientific groups and has been presented at conferences and seminars.

– During my time at Nexiona, I worked as a Data Engineer and Backend Developer for an industrial client in a manufacturing environment. – I developed a Python ETL infrastructure to ingest, transform, and process sensor data from different parts of the production chain, organizing it by time and section to provide accurate timeline analysis. – I implemented fault-tolerant mechanisms for sensor failures and delayed transmissions, built an intermediate NoSQL storage layer for data validation, and stored the processed data in InfluxDB as time-series data. – I also worked on anomaly and outlier detection and collaborated with frontend engineers and stakeholders to support the development of customized monitoring dashboards. – We followed Agile methodologies and used GitHub, GitLab, and Docker during development and testing.

– During my time at Volkswagen, I worked on the backend of touch screen applications for new car models, using Java and Oracle Java. – These applications were connected both to car sensors and to mobile-related services such as music and maps. – I also developed Python tools to speed up testing through touch screen simulation and to make test results clearer and easier to analyze. – We worked with Bitbucket, Jira, and Agile methodologies, and I collaborated closely with QA and Frontend teams to ensure the backend fit the needs of the full application.