| Project name | Description | Tools and technologies | Date |
|---|---|---|---|
| Paidwork sp. z o.o. | |||
| Document Video Registration | I developed a real-time machine learning model that identifies document types using camera data. If the document type was initially undetermined, the model dynamically retrained itself on new data in real-time. Captured data was transmitted to a server for further user verification, enhancing document recognition accuracy and enabling secure and efficient user authentication processes. This adaptive learning approach ensured continuous improvement and reliability of document identification. | - Python - Pytorch - OpenCV - Multiprocessing - Microsoft Azure | Apr 2024 Aug 2024 |
| Document Verification | I developed a machine learning model that utilizes NLP techniques to determine the type of a document based on its content. This model automatically processes the document and transfers the data to a database for further handling and storage. By leveraging natural language processing, the system accurately categorizes documents, streamlining data management processes and ensuring efficient document verification workflows. | - Hugging Face - Pytorch - SQLAlchemy - Pandas - Azure Databricks | Jan 2024 Mar 2024 |
| Minsk Electrotechnical Plant | |||
| Business Process Refactoring | As a team lead, I spearheaded the refactoring and reengineering of Product Data Management (PDM) business processes within the enterprise. The primary objective was to automate existing processes and eliminate paper-based document workflows. Through comprehensive restructuring efforts, we successfully revamped all processes, implemented electronic document management, and achieved full automation. This initiative streamlined operations, enhanced efficiency, and significantly reduced administrative overhead, marking a transformative shift towards modernized and agile business practices. | Team Lead role | Jan 2018 Sep 2023 |
| PDM System Implementation | I oversaw the implementation and ongoing support of a Product Data Management (PDM) system. This automated system facilitated the migration of all engineering and technical documentation into digital formats, ensuring immediate access to crucial documents. By implementing robust access controls, we bolstered system security, while the integrated platform significantly accelerated both the development and maintenance of documentation processes. This initiative streamlined operations, fortified data protection, and markedly enhanced productivity across engineering and document management workflows. | Team Lead role | Jan 2014 Sep 2023 |
| Return Prediction | I developed a machine learning model that predicted which type of product was most likely to require repairs in the current month based on historical data. This predictive model enabled proactive maintenance planning and resource allocation, ultimately reducing downtime and improving overall product reliability. | - Python - Scikit-learn - PostgreSQL - Pyspark - MongoDB | Dec 2022 Feb 2023 |
| Information Dashboards | I developed comprehensive information dashboards tailored for executive management. These dashboards provided a real-time overview of key production metrics, empowering leadership to monitor performance indicators and swiftly respond to any fluctuations or emerging issues. This initiative significantly improved decision-making processes and operational efficiency by ensuring that critical data was always accessible and actionable. | - Pandas - Power BI - Pyspark Databricks - MongoDB - Matplotlib - Snowflake | May 2021 Jan 2022 |
| Directory Export Tool | I developed a tool to extract directories from an outdated proprietary system and automatically populate these data into a new system. This seamless transition enabled the organization to smoothly adopt a new automated document management system. The tool facilitated efficient data migration, ensuring continuity of operations and enhancing overall workflow automation and efficiency in document handling processes. | - Python - Lxml - Pandas - PostgreSQL - SQLAlchemy | Sep 2020 Mar 2021 |
| Archive Transformation | I developed a tool that converts old scanned technical PDF documents into electronic formats. Following conversion, the data was processed and stored in an SQL database for subsequent use in ERP processes. This initiative streamlined access to historical data, enhanced data management efficiency, and integrated legacy documentation into modern digital workflows, thereby improving overall operational effectiveness and facilitating informed decision-making in ERP environments. | - Python - Tkinter - EasyOCR - Apache Airflow - PostgreSQL | Jan 2020 Nov 2020 |
| User Support Program | I developed a tool for logging user requests within our system. This program allowed for tracking the resolution of requests at any stage, provided management with insights into unresolved issues, and optimized the time of support staff. By streamlining the support process, the tool enhanced efficiency and ensured timely problem resolution, thereby improving overall user satisfaction and operational effectiveness. | - Python - Tkinter - Apache Airflow - PostgreSQL - Docker | Mar 2019 Jun 2019 |
| Production Order Automation | I participated in the development of a tool for generating production order forms. My responsibilities included integrating information from external sources into the application and exporting data from the completed forms to databases. This work enabled the automation of several critical business processes, significantly enhancing efficiency and streamlining the workflow in production order management. | - Apache Airflow - PostgreSQL - SQLAlchemy - Lxml - Pandas | Apr 2017 Mar 2019 |
| Technological Document Workflow | I automated the technological document workflow at the enterprise. I developed the necessary forms and approval processes for documentation. Essential directories were exported from the existing system. Once a technological document was created, all data were automatically transferred to the ERP system for further processing. This automation streamlined the workflow, ensuring seamless integration with the ERP system and enhancing overall efficiency in handling technological documentation. | - Intermech Search - Intermech workflow - Intermech TechCard - Intermech Imbase - PostgreSQL | Jan 2016 Aug 2017 |
| Engineering Document Workflow | I automated the engineering document workflow at the enterprise. The CAD system was standardized, and material directories and grades were exported. All engineers began working within a unified PDM system, with all documentation approval processes configured. This initiative reduced development and approval times, minimized errors during design, and enabled tracking of any document at various stages of production preparation. The automation significantly improved efficiency, accuracy, and transparency in the engineering workflow. | - Intermech Search - Intermech workflow - Intermech CadMech - Intermech Imbase - Autodesk Inventor | Dec 2014 Jan 2016 |
| Office Document Workflow | I transitioned the enterprise's office document workflow to a digital format. This included implementing access controls, creating a user-friendly archive search, and enabling management to conveniently assign tasks, monitor their progress, and obtain detailed reports on any office document. This digital transformation improved efficiency, enhanced oversight, and provided comprehensive insights into document management and task tracking. | - Intermech Search - Intermech workflow - Pymupdf - Pandas - Seaborn | Jan 2014 Feb 2015 |
| Pet-projects | |||
| Portfolio Website | I developed this portfolio website as a promotional tool using Flask. It is hosted on a remote virtual machine on Microsoft Azure and deployed using Docker. The website serves as a showcase of my skills and projects, providing a professional online presence that highlights my work and expertise. | - Python - Flask - Nginx - Certbot - Docker | Jul 2024 |
| Spark Pipeline | This educational project involves building a data processing pipeline in PySpark and subsequently training a model with automatic hyperparameter tuning. The project utilizes PySpark to construct a robust pipeline that handles data preprocessing, feature engineering, model training, and evaluation. Automatic hyperparameter tuning ensures optimal model performance without manual intervention. | - Python - Spark - PySpark MLlib | Dec 2023 |
| Kaggle Competition | In this Kaggle competition, I employed various techniques with basic machine learning models. I extensively utilized the scikit-learn library, implementing data processing pipelines, training multiple models with hyperparameter tuning, and selecting the optimal model through ensemble voting. This project sharpened my skills in data preprocessing, model selection, and ensemble learning techniques, demonstrating proficiency in leveraging machine learning for competitive data science challenges on Kaggle. | - Python - Scikit-learn - Pandas | Nov 2023 |
| Visa Slot Bot | I developed this project for personal use to secure a visa appointment. Available slots were quickly taken, and I didn't want to pay intermediaries. My program continuously monitored for available slots and sent notifications to my phone when slots opened up. This solution allowed me to efficiently secure an appointment without the need for third-party services. | - Python - Selenium - Telegram-bot | Aug 2023 |
| Website Monitoring | This was an educational project focused on tracking website visits with visualization in Grafana and notifications in Telegram. The solution was deployed in Docker, providing a comprehensive monitoring system that allowed for real-time tracking, insightful visualizations, and prompt notifications to ensure optimal website performance and user engagement. | - Prometheus - Alertmanager - Grafana - Docker - Telegram-bot | Nov 2022 |
| Data Preparation | This educational project focuses on data preparation, involving key steps such as data preprocessing, integration of data from various sources, and populating a database with prepared data. The project is implemented using Apache Airflow, which orchestrates and automates these data pipeline tasks. By leveraging Airflow, I ensured efficient data management and integration, enhancing my skills in workflow orchestration and data engineering practices. | - Python - Apache Airflow - PostgreSQL - Sqlalchemy - Docker | Aug 2022 Oct 2022 |
| Handwritten Text Recognition | I devised this educational project for myself to delve into the workings of a multilayer perceptron with backpropagation. Using only Numpy, I implemented a system for recognizing handwritten text without relying on other machine learning libraries. This project allowed me to gain a deeper understanding of neural networks and their applications in pattern recognition, reinforcing my proficiency in implementing machine learning algorithms from scratch. | - Python - Numpy - Tkinter - OpenCV | May 2021 Jun 2021 |
| iPhone Daily Planner | I undertook a personal project using Swift to develop a daily planner app for my iPhone. I implemented all the necessary functionalities, including reminders and widgets. This self-developed app allowed me to move away from the limitations of existing applications, significantly enhancing the convenience of using my iPhone and boosting my productivity. | - Swift - SwiftUI - WidgetKit - UserNotifications | Feb 2021 Apr 2021 |