M.Sc. Mechatronics & Robotics · Frankfurt University of Applied Sciences
Passionate about building complete systems: from bare-metal embedded firmware and real-time operating systems to computer vision pipelines and robot autonomy. I work across the full stack — STM32 firmware in C, ROS 2 nodes in Python, TinyML deployment with X-CUBE-AI, and custom PCB design.
Peripherals: DCMI · DMA · UART · I²C · SPI · CAN Bus
Tools: STM32CubeIDE · STM32CubeMX · X-CUBE-AI · MATLAB/Simulink · Autodesk Fusion / Eagle
Topics: Mobile robot control · Encoder odometry · PID control · Camera-based line following · TF2 · OccupancyGrid mapping
Topics: TinyML · MobileNet · Post-training quantisation (PTQ) · On-device inference · ST Model Zoo
STM32H7 · FreeRTOS · OV2640 · X-CUBE-AI · PySide6
Real-time edge vision system on STM32H743ZI. Classical CV pipeline (Otsu threshold, morphology, run-length CCL) and TinyML object counting (MobileNetV1-0.25, 96×96 RGB, 5 classes) — all inference on-chip. PySide6 dashboard for control, benchmarking and dataset capture. Full ML training pipeline (TensorFlow → TFLite PTQ → X-CUBE-AI deployment).
ROS 2 · Python · OpenCV · FreeRTOS · PID control · Dead-reckoning
Autonomous line-following and path-mapping for a Zumo robot. Camera-based line detection with Kalman filter, PID motor control over serial, encoder-based dead-reckoning odometry, TF2 broadcasting, OccupancyGrid mapping, and a PyQt5 live dashboard.
C++ · ATmega328P · Interrupt-driven encoder · PID
Ported the Pololu Zumo 32U4 encoder library to the ATmega328P (Arduino Uno). Software-based XOR replacement for quadrature decoding with attachInterrupt(), signed 32-bit atomic tick counters, and a discrete-time PID motor controller.
MATLAB/Simulink · CAN Bus · Vehicle Network Toolbox · DBC
MATLAB/Simulink simulation of CAN messages for a BMW E9x instrument cluster. Custom DBC file, signal generation, and hardware-in-the-loop testing via the Vehicle Network Toolbox.
STM32F4 · PCB Design · DRV8848 · INA138 · MATLAB PID Tuner
Custom PCB for a real-time PID demonstrator: STM32F4-Discovery as controller, DRV8848 H-bridge, INA138 current sensing, voltage regulation (12 V → 3.3 V), and hardware-adjustable Kp/Ki/Kd via potentiometers. PID parameters tuned with MATLAB PID Tuner.
📍 Frankfurt am Main, Germany