About Khalil
French
Native or bilingual
English
Conversational
Arabic
Fluent
Experience
- Université de Caen Normandie — Master 1IAResearch Project — Human Skeleton Reconstruction from LiDAR Point CloudsHEALTH AND WELLNESSNovember 2025 - June 2026 (7 months)Caen, FranceAnnual research project (team of 4) focused on reconstructing a 3D human skeleton from LiDAR data for movement analysis (rehabilitation, sports, 3D animation).Work performed:• Acquisition and processing of LiDAR point clouds (RS-Helios-16P sensor, ~28,000 pts/frame)• Cleaning pipeline: filtering, voxel downsampling, outlier removal• Geometric segmentation: RANSAC (plane removal) + DBSCAN (clustering)• Implementation of an L1-medial contraction 3D skeletonization algorithm (extraction of 15 anatomical joints)• Integration with MediaPipe for RGB/3D fusion• Temporal smoothing pipeline (Savitzky-Golay filter, interpolation)• Animation of a 3D avatar in Unity using Inverse KinematicsResult: a complete system capable of transforming a point cloud sequence into an exploitable animated skeleton, with a PyQt6 visualization interface (4 synchronized views).Tools: Python, Open3D, NumPy, SciPy, MediaPipe, Unity, PyQt6, scikit-image.
- Maison de l'Intelligence Artificielle de l'Université Mohamed PremierCompetition — Plastic Waste Detection by Computer Vision (YOLOv8)ENVIRONMENTALJune 2024 - June 2024Oujda, MoroccoParticipation in a Computer Vision competition focused on automatic detection of plastic waste.Work performed:• Dataset creation and annotation• Training a YOLOv8 model for real-time object detection• Hyperparameter optimization to improve model accuracy• Performance evaluation on test setResult obtained: mAP50 of 0.80, enabling reliable detection in real-world conditions.Tools: Python, YOLOv8 (Ultralytics), OpenCV, PyTorch.Github
- École Supérieure de Technologie OujdaMachine Learning Intern — Predictive AnalysisHEALTH AND WELLNESSApril 2024 - July 2024 (3 months)Oujda, MoroccoEnd-of-year internship for a Bachelor's degree focused on predictive analysis of depression using a real dataset of the Bangladeshi population.Missions performed:• Data cleaning and preparation (feature engineering, missing value handling)• Handling class imbalance with SMOTE• Comparison of 12 Machine Learning models (Logistic Regression, Random Forest, SVM, XGBoost, etc.)• Testing 7 variable selection methods (including Boruta)• Rigorous evaluation using cross-validationResult obtained: best model (Logistic Regression + Boruta) with 93% Accuracy, 0.98 AUC, and 92% F1-score.Tools: Python, scikit-learn, pandas, numpy, imbalanced-learn.Github :
Recommendations
Be the first to recommend Khalil
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Master's degree in Computer Science, AI and Human Factors trackUniversity of Caen Normandy CaenMaster informatique parcours IA et facteurs humains
- Professional Bachelor's degree in Decision InformaticsSchool of Technology2024Licence Professionnelle en Informatique décisionnelle