I'm an economist with a PhD from the University of Strasbourg, interested in the intersection of software engineering, programming languages, and reproducible data science.

My work focuses on building tools that make analytical software easier to develop, reproduce, compose, and maintain. I'm particularly interested in functional programming, declarative build systems, and applying software engineering principles to scientific computing.

I develop open-source software primarily in R, with a strong focus on reproducibility and developer tooling. Much of my work revolves around Nix, which I use to explore how software environments, dependencies, and computational workflows can be described explicitly and reproduced reliably.

Some of my projects include:

• rix — An R package that simplifies the creation of reproducible development environments using Nix.
• rixpress — A framework for building reproducible analytical pipelines with Nix, supporting workflows written in R, Python, and Julia.
• ryxpress — A Python port of {rixpress}.
• chronicler — An R package for instrumenting functions and recording execution details, inputs, outputs, and errors.
• cronista — A Python port of {chronicler}.
• T — An experimental, reproducibility-first domain-specific language for polyglot data science.
• Korppi — An offline-first text editor for asynchronous collaboration, built around patch-based document history.

Across these projects, I'm interested in making computational workflows more explicit, composable, and reliable, while reducing the gap between exploratory analysis and maintainable software.

My main interests include reproducible computing, functional programming, programming language design, and analytical pipeline orchestration. I'm particularly interested in the design of tools that make computational work reproducible by construction and explicit.

I enjoy exploring how ideas from software engineering can improve statistical computing: not just by making code run, but by making it easier to understand, test, share, and rebuild across machines and over time.

I write about R, statistics, software development, and reproducible analytical workflows on my blog.

I'm also the author of two books:

• Building Reproducible Analytical Pipelines with R — A practical guide to reproducible analytical workflows, combining R, Git, GitHub, Docker, and software engineering principles. Available to read online for free.
• Reproducible Polyglot Data Science — This book focuses on setting up polyglot data science projects using Nix and my bespoke DSL, T. Still WIP.
• Modern R with the tidyverse — A guide to modern R programming, covering the tidyverse, functional programming, statistical methods, and package development. Available to read online for free, but kinda dated now.

I regularly give talks and workshops on reproducible data science, Nix-powered development environments, and analytical pipeline engineering. These presentations combine practical demonstrations with broader discussions about reproducibility, software design, and reliable scientific computing.

I've presented at events including FOSDEM, R/Pharma, R Medicine, and at meetups from several organizations. You can explore my presentations through the repositories, slides, and recordings linked on my talks page.

I'm interested in pushing reproducible data science beyond environment management towards a more comprehensive approach to computational reproducibility. This includes reproducible builds, cross-language orchestration, functional composition, and programming language tools that make dependencies and execution explicit.

Ultimately, I want to build software that helps people spend less time managing fragile computational setups and more time doing meaningful analytical work.