Jeremi Wójcicki

I developed a web application with a back end to collect, store, and analyze energy data from power consumption sensors located at customers' locations. It included a machine learning random forest regression model to forecast future demand based on historical data and customer meta-data. I also created a front end to visualize data and predictions.

In this research work, I created a framework to mathematically model energy demand and processing rate of operations and technological processes in industrial machinery. It combined the use of analytical, simulation, and black-box models (numerical digital-twins).

My work included using regression methods for model identification and sophisticated optimization algorithms to minimize energy use and ensure the highest productivity. I was responsible for the delivery of this project, which being part of a larger European research project, had an allocated budget of €250,000.

I implemented an Eulerian motion magnification algorithm developed by MIT exploiting GPU acceleration, which allows efficient processing of large, high frame-rate video captures. The purpose is to extract micro motions invisible to the human eye for various industrial applications, e.g.:
- Intuitive operational deflection shapes visualization
- Dynamic analysis of micro-motions without using accelerometers