Deep dives into machines, minds and the people who build them.
Essays and hands-on guides by Max, AI architect, neuropsychologist and dive instructor. Long form, tested code, no hype.
- Entries logged
- 29
- Categories
- 5
- Now scanning
- Neurosciences
- Last dive
- 21 Nov
Latest dive
One long read, one repository, every number reproduced on a laptop.
-
From the dev environment to Kubernetes: the life of an ML model, explained to CI/CD engineers
A field guide for CI/CD engineers who have never trained a model: training and sweeps in a dev environment, a release pipeline to an artifact registry, a decoupled serving pipeline on Kubernetes or OpenShift, and why monitoring the model in production is run, not CI/CD. With code, MLflow screens and an open source multi project repository.
The chart
Each contour line is a category. Each sounding is one entry, named by its main idea. The orange route links the last six, oldest to newest.
Who is diving
Same person, three trades. They meet more often than you would think.
15 years in AI
I build AI in places where it can’t fail: defense, health, air-gapped networks. Before that, a master’s in neuropsychology and research at the CNRS. On weekends, I teach diving.
This blog is where the three meet: how machines learn, how minds work, and what both do to the people around them.
Clip me in
One email when a new dive is logged. No more than twice a month.