Geoinformatics Student @ Wrocław University of Science and Technology
I specialize in merging Computer Vision and GIS to build high-performance geospatial tools. I enjoy developing end-to-end systems — from fine-tuning Vision AI models and implementing graph algorithms to deploying full-stack web and desktop applications.
Vision AI | GIS | Python
An end-to-end system for objective urban space assessment using Vision Transformers (DINOv2).
- Developed an automated route planning module for data acquisition based on the Chinese Postman Problem algorithm.
- Built a training engine based on TrueSkill rankings and Optuna optimization.
- Automated image data inference
Python | Geopandas | ETL
Desktop application for rail connection optimization.
- Implemented Dijkstra’s algorithm on time-expanded graphs.
- Developed a module for calculating ticket prices based on actual track geometry
📍 Mapories
FastAPI | PostGIS | React | Leaflet
Web application for geolocating memories with focus on privacy and performance.
- Optimized spatial queries using bounding-box filtering in PostGIS.
- Implemented JWT authorization with token rotation (HTTP-only cookies) and multi-layered bot protection (reCAPTCHA, Honeypot).


