Satellite Object Detection
Detection of moving cars and airborne objects in satellite video using an edge-aware vision pipeline.
AI · Embedded Systems · Computer Vision
Mohamed Amine Mechlia
Embedded Deep Learning Engineer
— 01
About
Deep learning specialist building satellite and drone perception systems that operate reliably on edge hardware.
I design embedded AI pipelines for computer vision, 3D reconstruction, and real-time detection using Python, C++, TensorRT and hardware acceleration.
I work transparently in agile teams, share progress closely, and connect research to mission-ready deployments. Based in France, I also mentor kids through robotics education.
— 02
Skills
Python90%
Machine Learning75%
C75%
Computer Vision80%
C++80%
Embedded Systems80%
Deep Learning
PyTorch
TensorFlow
Keras
Computer Vision
OpenCV
MMDetection
Detectron2
YOLO
Data Science
NumPy
Pandas
Scikit-learn
Matplotlib
MLOps
ONNX
TensorRT
Docker
TorchScript
Geospatial
Rasterio
GDAL
SentinelHub
Albumentations
Edge AI
Jetson
OpenVINO
Vivado HLS
FPGA
— 03
Projects
Cloud Removal Module
Onboard cloud filtering for satellite imagery to reduce bandwidth and improve analytics accuracy.
GPS-Denied Drone Navigation
Onboard vision and SLAM for autonomous drones operating without GPS in challenging environments.
Embedded AI for Airborne Detection
Real-time airborne object detection and tracking for anti-drone applications on edge hardware.
Synthetic Aerial Imagery
Generated annotated aerial datasets with Blender and RenderDoc for training detection networks.
Real-Time Structure from Motion
Realtime 3D reconstruction for drone and satellite imagery using optimized visual SLAM and SfM pipelines.
ROS 2 Autonomous Driving Pipeline
Built a full autonomous driving stack in ROS 2 and Gazebo Ignition, including a tri-camera PilotNet deep learning model, PID control, and real-time perception.
Autonomous Robot Navigation — PACR
Designed a fully autonomous ROS 2 navigation stack integrating A* path planning, Pure Pursuit control, and an MDP-based task planner solved by value iteration — enabling a robot to pick, transform, and deliver objects continuously with zero manual input.
Projet ATLAS — Assistant hôtelier autonome (Robot Pepper)
Développement d'une architecture robotique complète sous ROS 2, ROS 1 (ros1_bridge) et Docker. Implémentation d'une solution VSLAM 2D (RTAB-Map) par caméra RGB-D et d'un planificateur A* sur grille (inflation d'obstacles, contrôle proportionnel lookahead). Intégration du moteur HRI QiChat et interface tablette pour la gestion de scénarios d'accueil et de guidage.
Fruit Fly Detector
Embedded computer vision for insect detection in intelligent traps with real-time response.
Face Landmark Detection
468-point face landmark system built with OpenCV, TensorFlow and MediaPipe, optimized for CPU.
ADAS FPGA Accelerator
Accelerated pedestrian detection on FPGA for advanced driver assistance systems.
TensorFlow Developer
Convolutional Neural Networks
Scrum Fundamentals Certified
Machine Learning Course
Meeting Yann LeCun at ICCV
Green'it Tunisia first prize
ICCV demo presentation
Neobotics robotics startup
Entrepreneurship workshop
Visit to ESA Noordwijk
SnT Reliability & Trust Hub
ICCV Paris 2023
VivaTech 2026 — Demo Day
VivaTech 2026 — Panel
Most Proud Of: Founding LEGOLAB
If you ask me what I am most proud of in my journey, it isn’t just the complex code I write today—it’s the spark I managed to ignite in the next generation.
At just 20 years old, while navigating the heavy workload of my Bachelor’s degree, I decided to channel my passion for technology into something bigger: I founded a robotics and technology academy for kids. Starting from scratch, I took on the dual roles of entrepreneur and educator, designing a hands-on curriculum tailored for young, curious minds.
Through interactive workshops, I introduced them to the fundamentals of programming and hardware using LEGO robotics, Micro:bit, Arduino, and Ubtech robots. Transforming abstract, complex engineering concepts into tangible, joyful discoveries for children taught me more about leadership, communication, and adaptability than any textbook ever could. It proved to me that true expertise lies in the ability to simplify the complex and inspire others—a mindset I carry into every engineering challenge I tackle today.
Most Proud Of: Co-Founder & Technical Lead – IQClass
Founded at the age of 24 during the final year of my Master's studies, IQClass is a tech startup delivering specialized enterprise solutions[cite: 2]. We drive digital transformation through Industrial IoT, Automation, and high-performance Web & Mobile services, while offering strategic IT Coaching and corporate training for companies modernizing their infrastructure.
— 04
Experience
2024 – Present
Freelancer
Computer Vision & AI Consultant
Providing independent consulting and development services for embedded AI, computer vision, and robotics projects. Designing custom solutions for satellite imagery analysis, drone perception, and real-time detection systems on edge hardware.
Mar 2026
OVHcloud
Computer Vision Intern
Developed an indoor localization and mapping system using Visual Odometry and Visual SLAM techniques for server room navigation in Croix, France. Designed real-time visual tracking algorithms for server localization in complex environments. Integrated ARKit-based augmented reality visualization and localization capabilities. Performed camera calibration, trajectory evaluation, system validation, and performance benchmarking. Conducted testing and debugging under real-world deployment conditions.
2024 – Sept 2025
TELNET Space Mars
Embedded Deep Learning Engineer
Developed object detection and segmentation algorithms for multispectral satellite imagery with hardware optimization, cloud removal, and anti-drone detection systems.
Nov 2021 – Feb 2024
Nurenda Technologies
Embedded Computer Vision Engineer
Built GPS-free drone navigation, embedded 3D reconstruction, synthetic aerial dataset pipelines, and robust segmentation/detection workflows.
Feb 2020 – Mar 2021
Micro-Optoelectronics & Nanostructures Lab
Research Intern
Implemented and accelerated a neural network on FPGA for pedestrian collision detection using Python, C/C++ and hardware design tools.
Sep 2018 – Present
LEGOLAB
Robotics Coach
Founded a robotics education club for kids and led workshops on programming and hands-on robotics with LEGO, micro:bit and Arduino.
Jun 2017 – Jul 2017
AVIONAV
Intern
Participated in early research on electric aircraft design while collaborating with French industry partners.
— 05
Education
2025 – 2026
Master 2 AI & Vision/Robotics
Faculty of Science and Technology Nancy
Advanced specialization in artificial intelligence, computer vision, and robotics systems.
2018 – 2021
Research Master in Electronics
ISIMM Monastir
Focused on microelectronics, instrumentation, and embedded systems for research applications.
2015 – 2018
Bachelor's Degree EEA
Faculty of Sciences Monastir
EEA — Électronique, Électrotechnique, Automatique.
— 06