Teufik – Python, FastAPI, Microsoft Azure

A district heating energy optimization service designed to help heat providers reduce production costs by leveraging the flexibility of heating systems. The solution combined advanced optimization techniques, including ADMM (Alternating Direction Method of Multipliers), with machine learning approaches such as Lasso regression for accurate heat load forecasting.

• Commercialized the project from its initial stage into an industrial-scale API service with automated deployments to the cloud platform;
• Restructured the microservices architecture to improve efficiency, scalability, and maintainability;
• Improved heat load forecasting accuracy by up to 10% by changing the approach to network heat load curve processing and switching to a linear regression model;
• Collaborated with researchers in the heat energy domain, gaining and applying domain-specific knowledge;
• Fostered knowledge sharing and the continuous development of software engineering skills within the team.

A bidding strategy library for industrial energy market participants, built on the Pyomo optimization framework. The solution models various industrial assets, such as wind turbines and steam engines, as well as multiple energy markets, including gas and day-ahead electricity markets, while integrating market price data to support optimization and bidding strategy development.

• Designed and established the project structure and optimization framework;
• Implemented multiple energy market models and industrial device models;
• Developed a reporting mechanism supporting both graphical outputs and scientific paper formats;
• Conducted knowledge-sharing sessions and training for domain researchers;
• Developed and applied domain knowledge in energy markets and optimization.

A data-feed pipeline built with Scrapy for collecting and processing vehicle data from multiple retail websites across the US. The project also included a Django-based MVC application that enabled manual matching and management of vehicle type data.

• Expanded and maintained a data feed pipeline to ensure reliable data collection and processing;
• Resolved web scraping tasks and addressed related technical issues;
• Designed and developed an MVC application from scratch for data matching, along with a custom administration portal;
• Reported project progress and technical updates to Product Owners and key stakeholders.

An administrative portal for securely accessing and managing patient data related to cell therapy, designed in compliance with HIPAA requirements. The project also involved developing and configuring a custom OAuth 2.0 authorization server to provide secure user authentication and access control.

• Set up and configured the OAuth 2.0 authorization server from scratch;
• Designed and developed the custom administrative web application.