Kyiv, Ukraine Right know Im consulting companies on how to digitally transform with Data Science, Machine Learning and Computer Vision

Kyiv, Ukraine This project required full cycle of data science activity including : data exploration, exploratory data analysis, feature engineering, training, deploying and audit machine learning pipelines. The main contibutions here were participating and supporting in all of the steps mentioned above, that resulted in first models deployment in a half a year after research started. Along this I've conducted a research for model improvement with classical approaches and NLP, support of ongoing process of deeper machine learning integration to meet customers demand and creating insights about product processes.

Kyiv, Ukraine The main activity here were PoCs. First PoC was localization of objects in specific state and localization them on map. The main contributions were developing Computer vision pipeline and backend for application. Computer vision pipeline consists of objects detection with CNN, multiple objects tracking part and anomaly detection. Backend was build in microservice fashion and support camera stream management adding, capturing and deleting and deployment of highly computationaly intensive part CNN on separate machine with GPU. Second PoC was connected with measurement of certain geometric features of an object. The main contribution was developing Computer vision pipeline. Ones consists of depth measurement with 3D camera, image segmentation with CNN, length and area measurement. For estimation of real values of geometric features supportive calibration pipeline was developed.

Lviv Region, Ukraine The main activities on this position were several proof-of-concepts related to object detection with deep learning and Computer vision in edge computing. Main activity for first PoC was adaptation of several CNN frameworks RFCN, Dense-RFCN, Deformable-RFCN, etc. to main computer vision pipeline. Therefore main contribution were speeding up process of model training and deployment, speeding up Computer vision pipeline with recent more faster CNN architectures. Second PoC was connected to extending Computer vision pipeline on set of Raspberry Pies. The main contibution were deploying aruco marker detector on Raspberry Pie and developing an API for interfacing with main application.

Nancy, France During this internship I've supported two professors from France and Mexico in their reseach and was the main contributor to codebase of the project. The key idea of the research was to explore self-organizing properties of Kohonen's map and its possible application to reliable computing in multiprocessing systems. Preliminary results got published as part of workshop WSOM+17. My main contribution was speeding up simulation and introducing some minor ideas to whole experiment.

Kyiv, Ukraine This was an internship in digital intergrated circuit department. I've leant basic programming concepts and specific for IC design and verification. The size of the projects varying from small finite state machine to large like MIPS core and USART.