Ahmed Krimi

Specialized in Robotics, ML/Computer Vision, and automotive software: - Led drone-based object detection/tracking by collecting aerial data, processing it, and fine-tuning vision models with transfer learning. Implemented these models in ROS2 for real-time use. - Built a GDPR-compliant LiDAR-based person detection system. - Developed a framework for autonomous driving validation in ROS2/Autosar Adaptive and C++, focusing on agile methods, clean code, and comprehensive documentation.

Indoor ultrasonic-based localization of a mobile robot using Time Difference of Arrival: - Designed a motion and sensor model, enabling tracking of robot movements within dynamic environments - Developed two recursive algorithms, using optimization and particle filters, to infer robot positions during motion - Developed two recursive algorithms, using optimization and particle filters, to track robot positions - Demonstrated the performance of the algorithms through simulations and real-world experiments

Project: Development of a Microelectronic Chip Defect Detection Tool with TensorFlow and Python Achievements: - Pre-processed the images using OpenCV - Created and labeled the dataset, including data augmentation techniques to avoid overfitting - Fine-tuned and evaluated a Faster R-CNN pre-trained model using the TensorFlow Object Detection API - Integrated the model into a GUI developed in Python to accurately identify defects and extract ID from the chips

Project: Design and implementation of a Nonlinear Model Predictive Controller (NMPC) Milestones: - Derived a global kinetic model to accurately represent the car's motion - Formulated the Non-Linear Model Predictive Controller (NMPC) optimization problem, incorporating constraints - Implemented and evaluated the NMPC in MATLAB, ensuring tracking of the generated trajectory

Project: Development of the "Chip Selection Tool" in VB.NET Features: - Comprehensive GUI - Processing and visualization of measurement data - Creation of pdf-reports with the filtering results - Statistical data analysis and outlier detection

Master Thesis: Indoor ultrasonic-based localization of a mobile robot using Ultrasounds andTime Difference of Arrival (Grade: 1.0) Key Courses: - Robot Localization and Mapping - Deep Learning - Numerical Optimization - Numerical Optimal Control - Computer Vision - Modeling and systems Identification - Microcontroller Techniques - Real-time operating systems (RTOS)

Two years of intensive preparation for the highly Competitive Examinations leading to the top Grandes Ecoles. Main courses : - Advanced Mathematics, Physics. - Engineering Sciences. - Computer science. - French philosophy.