Python ML Engineer (3.5y) · Computer Vision · Robotics-curious · Building small, practical tools
I’m a Python ML engineer with ~3.5 years of commercial experience, focused on applied ML and production-friendly Computer Vision. Currently building pet projects that help me practice, explore ideas, and stay sharp.
- ML/CV: PyTorch, TensorFlow, OpenCV, TensorRT, CUDA
- Backend: FastAPI, asyncio, multiprocessing
- Data & annotation: Label Studio (workflows, exports, automation)
- Ops: Docker, Linux, Git, CI/CD
- Robotics/sim: Gazebo, Webots, basic ROS 2
- Video: GStreamer
- End-to-end CV pipelines: data → annotation → training → export → deployment
- Semi-automated segmentation (SAM/FastSAM) to speed up labeling loops
- Simulation-based test scenarios for faster iteration and regression coverage
- On‑prem deployments: debugging, performance checks, hotfixes, data extraction
-
Kivy ML App — Kivy/KivyMD desktop app for secure image workflows + ML training (classification) + YOLO detection, with pytest integration tests.
https://github.com/KaMeLoTmArMoT/Kivy-App -
Explainable-AI-MRNet — Explainable deep learning on the MRNet knee MRI dataset (CAM + SHAP) with scripts for training, attribution, and visualization.
https://github.com/KaMeLoTmArMoT/Explainable-AI-MRNet -
ReadingsTracker — browser-only consumption tracker (CSV + charts + export).
https://github.com/KaMeLoTmArMoT/ReadingsTracker -
CardMatch — browser-only language-learning / matching game powered by CSV rows.
https://github.com/KaMeLoTmArMoT/CardMatch
A Novel Explainable AI Model for Medical Data Analysis — JAISCR 14(2), Mar 2024
License: CC BY‑NC‑ND 4.0 (Open Access).
Focus: explainable AI for medical imaging / high-dimensional data (hybrid ensemble).
Authors: Nataliya Shakhovska, Andrii Shebeko, Yarema Prykarpatskyy.
Links: Paper · PDF (offline) · Code
Development of the Architecture of Document Optical Character Recognition System — Herald (Tech. Sci.) 309(3), May 2022
License: CC BY 4.0.
Focus: document OCR pipeline (text detection/segmentation + recognition for Ukrainian/English).
Authors: Nataliya Shakhovska, Andrii Shebeko.
Links: Paper · PDF (offline)