software

Projects I’m currently working on:

• rustineers: A collection of crates to learn Rust with applied sciences. I try to comment and explain my journey in a small book that goes with it.
• plaid: An open-source project developped by my colleagues.
• cppxplorers: A C++ mono-repository playground. There’s a book I try to fill as well.
• kernax: A small JAX-based package that implements kernel methods. I’ve been refactoring that library a bit.
• pbnn: A code that benchmarks UQ methods for neural networks in JAX and PyTorch. I’m cleaning this up as well.
• Python lecture notes: Some lecture notes for a lecture I give about implementing algorithms in Python. I’m trying out marp.
• Marimo notebooks: Some notebooks that go along with the lecture notes. Trying to make nice visual examples.

Experimenting:

• uv python monorepo: I made a template for a python mono-repository setup with uv. I added all the tools I like to use right now. It can easily be used as a standard repository too.
• endurance: An example of python playground based on the above template.

Older stuff:

• TrilinosUQComp: A 3D parallel finite element solver I made during my PhD thesis. It relies on Trilinos and good old Epetra (which is deprecated now). It was my first C++ project, it has several flaws but I was happy with it! Nowadays I would probably use Spack and rewrite the finite element solvers with Kokkos/Tpetra.
• BaTorch: A code for benchmarking Bayesian neural networks with PyTorch and Hydra We implemented several Bayesian methods for deep neural networks, and kernel-based distances stuff like the maximum mean discrepancy and the kernelized Stein discrepancy. It is outdated but I plan to use Hydra again in another project.