selected_works:
Sign_Language
Python computer vision model for live hand-sign detection
H2O_AutoML
Titanic survival prediction with automated ML workflows
Heart_Disease
Decision tree classifier for clinical risk prediction
Digit_SVM
Handwritten digit recognition using support vector machines
( behind the outcome )
Good questions.
Better systems.
The interesting work happens between an idea and its next iteration.
Observe.
Start with the question. Look closely at the data and the problem it represents.
Experiment.
Make an idea testable. Build a small version and let its behavior guide the next step.
Question it.
Inspect the failures, not just the result. Refine what matters and test again.
( applied intelligence / 001-004 )
I build
systems
that see
and learn
from data
to make
useful
products.
Frame. Train. Prove.
Problem / experiment / evidence
( university education )
Academic
foundations.
Université
Paris-Saclay / UVSQ
Master AMIS
Algorithmique et Modélisation à l'Interface des Sciences
University of
Khemis Miliana
( tools into practice )
The working
set.
From an input to something useful.
The tools change with the problem. The process connects them.
- 01ExamplesPython / datasets
- 02FeaturesOpenCV / preprocessing
- 03TrainingPyTorch / model fitting
- 04EvaluationScikit-learn / held-out data
Examples become features. Evidence decides what works.
Turn raw examples into a model, then check what it has learned.