Oleksandr – Python, PyTorch, Machine learning

The project is an FMCG sales management system that helps to plan, execute, and monitor your sales.

• Made product architecture changes that improved the accuracy;
• Participated in product future development planning;
• Significantly decreased system deployment pipeline time;
• Was in charge of researching, designing, and implementing Machine Learning algorithms for different tasks;
• Made an automation of manual processes and optimization data pipelines, interpreting it in an interactive way;
• Provided technology and scientific consulting performance;
• Was communicating with business and technical stakeholders to identify solution requirements, business goals, and the expected outcome;
• Made R&D and feasibility studies to share results and methodologies;
• Was delivering rapid prototypes and proof of concepts that demonstrate real value.

This is photo editing software with a simple interface and deep learning models under the hood.

• Developed three photo editing features;
• Built an infrastructure for conducting Machine Learning experiments;
• Created custom model architecture for specific tasks;
• Developed new and improving existing machine learning models used in the product;
• Solved unique problems that do not have ready-made solutions;
• Was responsible for data collection, training, evaluation, and making iterative improvements;
• Designed and implemented complex algorithms.

KYC system for financial institutions that helps to pass all the checks online.

• Built the Machine Learning part of the project from scratch;
• Participated in creating the initial project architecture;
• Created custom algorithm based on OCR;
• Made research and development of KYC solutions;
• Contributed to the product concept planning from a data science and backend perspective;
• Participated in DS activities planning;
• Was in charge of Ad-hoc research, design, and implementation algorithms for KYC use cases.

The project was the system for plane surface inspection and defect detection.

• Changed product architecture and improved accuracy;
• Implemented models from research papers;
• Got insight into business problem, understanding the opportunity and value of analytical models for the customer;
• Was in charge of collecting, transforming and preprocessing raw data to prepare it for analysis;
• Was deriving descriptive statistics out of the preprocessed data;
• Built statistical and probabilistic models;
• Designed, developed, trained, and tested data mining, machine learning and artificial intelligence and algorithms;
• Provided comparative research on different algorithms and models;
• Implemented the model in a form that can be easily used by engineers, documenting its interfaces;
• Delivered the model in a form that can be easily deployable and maintained.