Freelancer profile translated to English.

Description

Freelance in **Computer Vision**, **Data Annotation**, and **Mobile Robotics**, I assist companies in creating quality datasets and developing reliable AI and robotic systems.

I have solid experience in image and video annotation **(CVAT)**: detection, tracking, 2D semantic segmentation, as well as annotation of 3D LiDAR point clouds for complex scene segmentation and understanding. I participated in the construction of AI models (YOLO, detection, anti-fraud, object recognition) used in production.

In mobile robotics, I worked on advanced stacks: ROS2, Nav2, SLAM 2D/3D, OctoMap, 3D mapping, trajectory planning, obstacle avoidance, sensor fusion (3D LiDAR, 3D cameras, IMU, GPS, tachymeter), as well as embedded development (STM32).

👉 I combine rigor in annotation, expertise in 2D/3D perception, and mastery of autonomous systems to deliver effective, actionable AI and robotic solutions adapted to field needs.

Languages

English
Native or bilingual
French
Fluent

Workplace preferences

Can work on-site

Paris (up to 50km), Lyon (up to 50km), Toulouse (up to 50km), Marseille (up to 50km)

Mo-Ka Paris
Freelancer
TECH
January 2025 - July 2025 (6 months)
Paris, France
Mission:Data Annotation & AI Model Training Services – Startup specializing in computer vision

Role & responsibilities:

- Collaborate with the technical team to optimize datasets, improve annotation quality, and validate the performance of trained models.
CVAT Python YOLO excel Linux
Segula Matra Automotive Brognard,
R&I Engineer Intern in Robotics, Mechatronics and Embedded Systems
AUTOMOBILE
April 2024 - October 2024 (6 months)
France
Theme: Holonomic terrestrial robots, using 3D mapping for trajectory management

Role & responsibilities:

- Review all work done on the omnidirectional robot via ROS2.
- Analyze existing literature by comparing cutting-edge approaches and the most suitable sensors for developing 3D mapping to visualize the robot's environment.
- Create a 3D voxelized environment by implementing a 3D OctoMap with a 3D LiDAR.
- Generate static and dynamic transformations (TF2). Filter the ground and the robot's altitude to prevent their subsequent projection into a 2D map.
- Project the filtered 3D voxels into a 2D occupancy grid map and generate a usable map.
- Calibrate the IMU in static mode to improve data accuracy.
- Enhance robot localization by integrating an Extended Kalman Filter (EKF), fusing odometry and IMU data.
- Configure the SmacPlannerLattice plugin as a planning server in Nav2.
- Integrate the projected 2D map as a Custom Costmap plugin in Nav2.
ROS C++ Sensors STM32 Sensor Fusion
Laboratoire de Recherche en Robotique, Parallélisme et Systèmes Embarqués (LRPE) à l'USTHB Algiers,
Robotics and Embedded Systems Engineer Intern
AVIATION AND AEROSPACE
March 2020 - October 2020 (7 months)
Algeria
Theme: Detection and identification of a specific person using Deep Learning CNN for Real-Time drone applications

Role & responsibilities:
- Control drone stability with Sliding Mode and Backstepping Control, possibly analyzing the obtained results.
- Prepare the software architecture and ensure ideal compatibility under ROS.
- Efficient and robust implementation of a Multi-Layer Perceptron (MLP) for tracking.
- Implementation of a YOLO-v3 tiny algorithm for visual detection/identification of the target.
- Simulate drone behavior with Gazebo Simulator and conduct real-time experiments to improve performance.
YOLO ROS Gazebo drone CNN