Ivan Bukač

- Prototyped firmware for a new feature on an active product, enabling Ethernet communication between an STM32 device and an SPI-to-Ethernet chip - Worked in C++ to analyze and integrate solutions into the existing codebase. Worked closely with senior engineers to ensure compatiblity. - Used company developed RTOS to manage tasks. Used Git for version control. - Verified SPI transaction integrity via oscilloscope analysis. Skills: C/C++, STM32, Git, RTOS, Oscilloscpe debugging

- Developed and implemented a Kalman filter for real-time vehicle motion tracking. - Employed a neural network to perform object detection on LiDAR point cloud data, inte- grating the results into the tracking system. - Built the entire solution in Python - Used Git to communicate and share code with a work partner Skills: Python, PyTorch, Git, Computer Vision, Kalman Filtering

- Researched a new company product for remote monitoring and gate control based on lowe power STM32 processor. Assigned to start development based on prior system architecture. - Optimized and reduced power usage by 50% by debugging and refining deep sleep behaviour. - Utilized LoRaWAN and FreeRTOS to infer gate states and detect physical movement. - Worked closely with the hardware team to better understand the device power usage. Skills: C, STM32, Git, FreeRTOS, Power consumption optimisation

- Studied in an international academic environment (6 months), improved English proficiency, developed strong collaboration skills, and learned how to easily integrate into Swedish workplaces. - Learned about System Engineering, Sattelite Communications and Deep Learning

- Participated in a university-led research project focused on humanitarian demining technologies. - Contributed to hidden object number estimation algorithm which later formed the basis for my Master thesis.

- Thesis title: Algorithms for estimating number of hidden objects from pulsed metal detector data - Developed and tested evolutionary and nonlinear optimization algorithms in MATLAB for estimating hidden objects in demining scenarios. - Built a simulation environment using dipole approximation and validated performance through lab measurements

- Thesis title: Mathematical modeling of consensus in information geo- metry - Modeled consensus formation via opinion dynamics using Python, expo- nential family distributions, and differential equations.