Julien Girard-Satabin (them)

@[email protected]
110 Followers
139 Following
1.1K Posts

Academic on Trustworthy AI, with a hunch for formal methods. Interpretable AI, formal verification of artificial intelligence systems, and so on.
Developper of the CAISAR platform https://www.caisar-platform.com/

Also, I cook when I'm not doing any OCaml. I look for computing for emancipation, not control.

My opinions are mine alone, no endorsement from my employer.

All my posts are under CC BY-NC-SA: non commercial use and attribution. Don't feed my thoughts to an AI, thanks.

Research interestsTrustworthy Artifical Intelligence, Formal Methods, doing research well
Personal websitehttps://julien.girard-satabin.fr/
Research grouphttps://list.cea.fr/en/digital-trust-building-a-more-secure-digital-world/
ORCIDhttps://orcid.org/0000-0001-6374-3694
Julien Girard-Satabin (them)Feb 21
spoune 🕯️Feb 20
Deux soupapes sous un arbre Ă  came
Julien Girard-Satabin (them)Feb 6
mon étudiante qui réussit mieux mon TP que moi, c'est signe que j'ai bien fait mon travail ou que je manquais de café?
Julien Girard-Satabin (them)Dec 9
Undone Computer ScienceJul 17, 2025

Please help us spread the word about Undone CS 2026: the second conference on Undone science in Computer science

• A conference in computer science to pause and reflect on the epistemological and ethical dimensions of the field, through the concept of undone science

• Organised by computer scientists, for colleagues from across the field, but also social scientists, philosophers and historians of science, and other scholars, working on computer science

• Luxembourg, 23-25th March 2026 (hybrid)
• Calling for short talk proposals (1-3 pages abstracts)
• Submission deadline: October 16, 2025 (anywhere on Earth) (extended deadline)
• Post-proceedings model: we will send a call for full papers after the conference
• Some travel funding available for speakers

Full call for presentations at https://www.undonecs.org/2026/cfp.html

More info at https://www.undonecs.org/2026/

Picture: “Dusk in Luxembourg Grund” by Tristan Schmurr (modified), CC BY 2.0

Show threadJulien Girard-Satabin (them)Dec 2
@t_viard nous a présenté un super travail de science réflexive sur la production de savoir (et sa légitimisation) sur l'impact environnemental de l'IA
Julien Girard-Satabin (them)Dec 2
ben c'était top en fait d'accueillir @t_viard aux journées thématiques que je co-organise !
Julien Girard-Satabin (them)Nov 27

Oh, au fait, je co-organise une journée sur la qualité opérationnelle des programmes à base d'IA. Ça sera ce mardi 2 décembre, sur le campus de @centralesupelec.bsky.social de 9h à 16h30.

Tous les détails, y compris l'inscription, sont disponibles ici https://gt-iasiv.fr/journee-thematique/

Journée thématique

C’est avec grand plaisir que nous annonçons la première journée thématique du groupe de travail IASIV. Le fil directeur de la journée sera “Qualité opérationnelle des IA”. Lieu et modalités pratiques Elle aura lieu le mardi 2 décembre, de 9h30 à 16h30, à CentraleSupélec, sur le campus de Saclay, île-de-france. Notez dès à présent la date dans vos agendas! La journée comptera des présentations longues d’expert·es issus du monde académique et industriel. Des pauses permettront des échanges que nous espérons riches autours de posters. Un repas format buffet sera fourni.

GT IASIV
Julien Girard-Satabin (them)Nov 19
FMASWorkshopNov 19
The last paper of this session, "The CAISAR Platform: Extending the Reach of Machine Learning Specification and Verification", is an #iFM paper and is being presented by @rhapsodos
Julien Girard-Satabin (them)Nov 19
Chèvre MousseNov 19

git pull
git Ă©charpe
git bonnet
git gants

erreur : fusion impossible (température trop basse)

Show threadJulien Girard-Satabin (them)Nov 19

I think this is a lesson that arguments like "this language is safest" (that I often hear about Rust, but can be found in any PL conference) have limits.

A program does not lives in the void. It is written by humans (who firmly believed that error handling was not necessary at the time, probably for sensible reasons).

The database monitoring engine is interfaced within a wider scope, which we as developers are forced to make assumption on. It turns out that some assumptions turned out wrong.

Julien Girard-Satabin (them)Nov 19
Oh wow @FMASWorkshop is here ???? incr