Meet the AI cell
We run the AI track: sessions, projects, and demos that have to work on stage. These are the people behind it.
-
El-baydaouy Amine
Head of Cell
-
Nouhad Oulkaid
Vice Head of Cell
-
Mazouar Adam
Course maker & mentor
-
Abdelaziz El Ouarary
Hackathons
The machine learning principle
Some problems have no clean algorithm. You cannot write a reliable list of if-statements that separates dogs from cats in every photo: fur, pose, lighting, and breed keep changing from picture to picture. So we stop hard-coding the answer. We build a framework that learns the job from labeled examples.
Labeled examples
Many photos. One number each.








No hand-written rule for fur, ears, or breed. Pixels in, label out: the model learns the mapping from examples.
Learn, do not invent every rule
Pixels in. Label out. Train on many examples until the model finds the mapping on its own. That is the core of machine learning, and the rest of the year is how to do it well.
What this looks like in the wild
Three open-source systems that learn from data: vision, speech, and images from text. Open a repo, skim the README, then come back ready for Week 2.
We do not hand-write every answer. We train systems that learn. The year ahead is how.


