🤖 Hi, I'm Julien Dubuc!

Engineering Student @ SeaTech | Mechatronics, Robotics & Embedded AI

I am an engineering student at SeaTech, passionate about the intersection of software and hardware. I thrive on complex challenges. From developing Python algorithms for AI to designing embedded C/ROS systems, I am driven by the process of turning a complex idea into a functional, intelligent machine.

• 🔭 Current Focus: Mechatronics, Robotics, and Embedded AI.
• 🎯 Looking for: Internships or opportunities to apply my skills in real-world environments.

• Tech Stack: Python, Computer Vision (OpenCV), YOLOv8, `Qualisys
• 🎯 Challenge: Develop a low-cost, markerless alternative to expensive industrial motion capture systems (Qualisys) to track the 3D trajectory of an underwater robot in a test basin.
• 🛠️ Solution: Built a complete Python pipeline. Trained a YOLOv8 model for robust object detection in a distorted aquatic environment. Implemented multi-camera calibration (OpenCV, ArUco) and a stereoscopic 3D triangulation algorithm using the least squares method.
• 📈 Impact: Achieved a tracking precision of ~4.7 cm at a 3-meter distance, validated against the Qualisys ground truth. The project was highly praised by the COSMER lab jury for its technical viability and commercial potential.

• Tech Stack: ROS, Python, Gazebo
• 🎯 Challenge: Enable a TurtleBot to navigate autonomously in an unknown environment while safely avoiding obstacles.
• 🛠️ Solution: Developed a navigation stack using ROS and Python within a Gazebo simulation. Implemented a robust State Machine processing Lidar data for dynamic path planning and managing physical collisions via bumper sensors.
• 📈 Impact: Achieved reliable autonomous movement and collision avoidance, demonstrating strong proficiency in standard robotics frameworks.

• Tech Stack: C, MPLAB X, Microchip/ADC
• 🎯 Challenge: Build a low-level, collision-free control system for a physical autonomous mobile robot.
• 🛠️ Solution: Programmed an embedded control architecture in C using MPLAB X. Configured Microchip peripherals (ADC, Timers) and managed hardware interrupts. Designed an obstacle avoidance algorithm aggregating data from 5 telemeter sensors.
• 📈 Impact: Delivered a fully functional embedded system capable of real-time sensor processing and autonomous spatial awareness.

• Tech Stack: MATLAB, App Designer (IHM)
• 🎯 Challenge: Create a tool to predict and analyze the complex physical behavior of various vehicles under different driving conditions.
• 🛠️ Solution: Designed a custom simulation application using MATLAB App Designer. Implemented mathematical physics models to calculate trajectory, lateral acceleration, yaw rate, and drift across multiple scenarios (dry, wet, ice).
• 📈 Impact: Provided a scalable UI to simulate rollover risks and handling characteristics for city cars, SUVs, and hovercrafts.

• Tech Stack: Python, Machine Learning
• 🎯 Challenge: Build an AI capable of mastering a board game from scratch without pre-programmed strategies.
• 🛠️ Solution: Coded a Hexapawn game engine in Python and implemented a reinforcement learning algorithm. The AI systematically recorded winning states and pruned losing branches over multiple iterations.
• 📈 Impact: Achieved a model capable of perfect play. Successfully scaled the board size and generated a complete game tree for instant win-state calculation.

🌐 Explore my other projects (Image Processing, Breakwaters Simulation, Business Strategy) directly on my Interactive Portfolio.