Bartek Kuźma

• Data Pipeline Development: Designed Python pipelines for processing football data from multiple sources, enabling comprehensive opposition analysis. Integrated provider data and utilized web scraping to enrich datasets for data-driven decision-making. • Database Management: Developed workflows to organize match and player statistics for coaching staff and analysts. Implemented solutions ensuring data integrity across operations. • Analytics Tools: Built interactive dashboards with Streamlit and HTML for visualizing performance metrics. Maintained KPI system to evaluate performance against benchmarks. • Opposition Analysis: Prepared data-driven video compilations of opponents to identify tactical patterns. Collaborated with coaches to develop targeted analysis packages combining statistics with video evidence.

• Content Matching: Developed and implemented algorithms to match movie content from various providers, utilizing IMDb and Wikipedia data to create an effective popularity ranking system.

• Multimodal Emotion Recognition System: Led the development of a system to recognize emotions across multiple languages and modalities, improving its accuracy and robustness through comprehensive data augmentation and rigorous testing. • Content Matching: Developed and implemented algorithms to match movie content from various providers, utilizing IMDb and Wikipedia data to create an effective popularity ranking system.

• IVR POC System: Engineered an Interactive Voice Response (IVR) proof of concept, enhancing the system’s ability to accurately recognize and respond to user queries by creating specialized entity matchers. • Multimodal QA Pipeline: Contributed to the design and development of a multimodal question-answering pipeline for the SemEval competition, focusing on pipeline optimization and case handling, resulting in significant performance improvements.

• Validation of credit risk models (PD/EAD) • Preparation of detailed summary validation reports describing outcomes of performed analyses, providing expertise and relevant recommendations to Credit Policy Committee • Models performance analysis using SAS EG