Zacharie Buisson co-built the winning solution with me. Every number below is measured, and every notebook is stored with the outputs that produced it.

Disclosure: I am the author and maintainer of AutoCarver, an open-source (MIT) library. Read the library sections with that in mind. Every number can be checked from the repository and datasets linked below. All images are by the author, generated from the challenge data (Etalab Licence Ouverte 2.0).

Table of contents

The finish line, first

1. The problem

2. The winning recipe (2025) 2.1 Carving the features first

2.2 Frequency: a weighted multiclass model

2.3 Severity and tuning

3. What a year of releases actually bought 3.1 Multiprocessing: the slow step got fast

3.2 An ordinal target

3.3 Did the new version win by more?

3.4 Three ways to bin

4. Takeaways

Everything behind this article

Setup: exactly what was measured

References

···

The finish line, first

In 2025, we took first place in an insurance hackathon.…