I'm a software developer specialized in AI/ML and computer vision. I have over two years of commercial experience and around five years of academic experience in related fields.
I hold a degree in Mechatronics (specialization: Photonics) from Warsaw University of Technology and I'm currently completing a Master's in Information Technology (also at WUT). During my studies, I focus on AI (strong in computer vision) and I’m widening into other AI/DS tracks to tackle diverse problem domains.
This repository showcases a selection of non-commercial projects I’ve worked on during my academic and personal development. The projects are ordered (in my opinion) from the most relevant and technically impactful to more exploratory ones.
| ID | Name | Brief | Technology |
|---|---|---|---|
| 01 | SfM Reconstruction | Sparse Structure-from-Motion with extra geometric hints (approx. pose/depth/intrinsics) + GUI; reconstructs 3D point cloud from one/two images and masks. | C++17, OpenCV, PCL, Qt, CMake |
| 02 | HES Detection | Handwritten Electronic Schematics: synthetic data generator + object detection (components) + connection parsing into a matrix. | Python, PyTorch (YOLOv8, Faster R-CNN), OpenCV; C++ (generator, nlohmann/json), COCO JSON |
| 03 | RNN Raincast | Nowcasting on SEVIR IR069: ConvRNN / ConvLSTM / Peephole-ConvLSTM with a lightweight head; demo sample, checkpoint and notebook included. | Python, PyTorch Lightning, Hydra, NumPy, Matplotlib, Weights&Biases (opt.) |
| 04 | Mouse Anomalies | User verification via mouse dynamics (Balabit): segmentation → 37 features → classic ML for anomaly detection and supervised classification. | Python, scikit-learn, XGBoost, SVM, Isolation Forest, PCA, SMOTE, Pandas/NumPy |