I architect and develop packages for research and education, bridging the gap between rigorous science and creative arts. My work focuses on building reproducible workflows and intuitive tools for ecology, mineral resources, agriculture and hydrology.

I am passionate about crafting tools for data science.

• nuee — A multivariate data analysis Python package for numerical ecology, designed for high-performance scientific computing.
• tangent/ds — A generic JavaScript package for data science, bringing "Tidy" principles to the browser (Statistics, ML, Plotting). Check out my new course on Observable.
• meandre — A differentiable, experimental hydrology model built as PINN (physically-induced neural network), written to be akin to perturbations: once trained, changing parameters allow to forecast the effects of land/water management on the water regime. Still experimental, but already got promising results.
• jmonlabs — An organization dedicated to algorithmic music composition where I develop the jmon declarative format and jmon/algo.
• seekwellpandas — A Python utility extending pandas with consistent tidy methods.

I also maintain educational & community resources.

• [In French] Analyse et modélisation d'agroécosystèmes avec R — Foundational course material on R for ecologists. (See the active fork here).
• [In French] Introduction au calcul numérique avec Python — A primer on scientific Python for engineers.
• DJalgo's User Guide — A comprehensive guide on soundifying mathematics (fractals, genetic algorithms) in the browser.

The audio streaming frontend Dustic is a hobby project available at https://dustic.app, and can be installed everywhere as PWA. It plays audio from archive.org and the FunkWhale network. I built it as a privacy-first gateway to discovering unusual and intriguing music, audiobooks, and podcasts.

Professional background: Senior Engineer & Ph.D. in Data Science / Ecological Engineering.