Roman Hrabkov

- Engineered a custom dataset for object detection by sourcing, selecting, and annotating frames from raw drone video footage, creating a challenging real-world benchmark. - Trained a YOLOv8 model on this dataset and performed a detailed performance analysis, identifying the impact of real-world data challenges (e.g., class imbalance) on model metrics. - Achieved a final mAP of 0.575, with the model successfully learning to distinguish key objects like 'IFV' and 'car' with high precision.

Participated in a practicum centered on software development using modern computer vision technologies, where I worked on detecting road lanes with OpenCV, recognizing humans in images using the Dlib library, and implementing object prediction and recognition in photos and videos with ImageAI and OpenCV.

Year-long fellowship program - Completed an intensive fellowship program (96 seminar hours) covering economic theory, with a focus on topics like cryptocurrencies and blockchain. - Developed practical skills in making complex information accessible by authoring analytical articles for a broad aud

I transferred from Kyiv Polytechnic Institute to Khmelnytskyi National University due to family circumstances that required me to relocate. This transition has provided me with more opportunities for informal development, particularly in volunteering and economics. While continuing my formal educat