Matúš Boleček

Aarhus, Denmark · hi@matusbolecek.com · matus-bolecek

Education

Aarhus University
BSc. in Data Science
2025 – Present
Gymnázium J. G. Tajovského
High School Diploma — 1.0 GPA (equivalent to 12.0/12.0)
2021 – 2025
  • Focus areas: Mathematics, Computer Science, Social Studies

Experience

AI Team Member
Slovak Debate Association
8/2025 – Present
  • Developed a grade-validation layer for LLM-based answer scoring in a national competition, automatically routing uncertain grades to human review
  • Achieved a ~50% reduction in grading error compared to the previous solution, validated on a test set
Debate Team Coach
Gymnázium J. G. Tajovského
9/2024 – 6/2025
  • Coached 30+ students in argumentation, research, and public speaking
  • Organized training sessions and coordinated tournament participation
  • Led school debate team to Slovak Debate League finals
Basketball Referee
Slovak Basketball Association
9/2023 – 4/2024
  • Officiated regional youth basketball games
  • Demonstrated decision-making under pressure and conflict resolution

Research & Projects

Generative AI as a Tool for Modern Education
SOČ – Student Research Competition (55 pages)
2025
  • 1st Place National Level – Pedagogy-Psychology-Sociology Category View project
  • Analyzed AI applications in educational settings; built a working LLM pipeline for AI-assisted student portfolio generation GitHub GitHub
Regional Aspects of Electoral Behavior in Slovakia
Geography Olympiad (153 pages)
2025
  • 3rd Place National Level + Michal Zaťko Award for best paper View project
  • Examined and visualized correlations between socioeconomic factors and electoral support across regions
Internet Usage Analysis in Slovakia
European Statistics Competition
2025
  • 3rd Place National Level View project
  • Statistical analysis of digital adoption trends 2019–2024
Otherness in Schools
SOČ – Student Research Competition (46 pages)
2024
  • 1st Place National Level + UKF Rector's Award View project
  • Survey-based research on discrimination perception by Slovak students (n = 1,072)
  • Provided recommendations for reducing negative attitudes in schools

Skills

Programming: Python, Git, R / RStudio, SQL
Libraries & Frameworks: Pandas, NumPy, Scikit-learn, XGBoost, SQLite, OpenAI API, Streamlit
Languages: Slovak (Native), English C2 (CAE), German B1 (DSD 1 PRO), Danish (Beginner: Modul 1)
Research & Analysis: Quantitative & Qualitative Methods, MS Power BI, Statistical Analysis, QGIS, Excel, Data Visualization, Survey Design, Academic Writing
Other: Linux , Team Leadership, Public Speaking, Mentoring, Class B Driver's License
Interests: Music Production (FL Studio), Audio Engineering, Photography (Adobe Ps + Lr)

Technical Projects

Nitor Energy Intraday Trading Competition GitHub GitHub
2026
  • First experience with ML in intraday energy price prediction
  • Placed top 10 based on RMSE on the public leaderboard
Interactive Election Map Pipeline & Dashboard GitHub GitHub
2026
  • Built an interactive Power BI dashboard featuring TopoJSON mapping, automated demographic correlations, and time-series analysis
  • Engineered an automated Python (Pandas) data pipeline to extract, clean, and unify over 20 years of historical election and demographic data
Bscripts – Music Production Automation Pipeline GitHub GitHub
2025 – 2026
  • Built a modular pipeline for music releases: file organization tracked in SQLite, automatic Dropbox upload, and CSV generation for SocialChamp to handle YouTube scheduling
  • Reduced release time by 80% (from 20 minutes to 2), with gains compounding significantly in bulk
AUHack 2026 – Energy Market Dashboard GitHub GitHub
2026
  • 48-hour hackathon project; built an automated data pipeline to process weather and spot-price data
  • Developed an XGBoost multi-regressor to forecast European electricity prices and LightGBM micro-models to simulate grid shocks, deployed via FastAPI and integrated with a React dashboard
Audio Loop Filtering Tool GitHub GitHub
2026
  • Used an audio embedding model to automatically rank and filter music loops against a trained profile
  • Created to solve the problem of overfilling inboxes and time-consuming manual sorting

Additional Awards & Achievements

Last updated April 2026