Data Science Student | Machine Learning & Predictive Analytics
CNN, LSTM, Transfer Learning
U-Net, MobileNet, Segmentation
BERT, Transformers, Whisper
Pandas, NumPy, Scikit-learn
PostgreSQL, Snowflake, SQL
Streamlit, Matplotlib, Seaborn
Azure, Docker, CI/CD
NVIDIA cuOpt, Route Planning
Mixed-methods research for Digiwerkplaats with 178 survey respondents and 7 interviews. Information quality identified as key driver of chatbot satisfaction.
Complete computer vision system for NPEC with U-Net segmentation, root tracking, and cloud deployment. Achieved 11% SMAPE score for accurate root length prediction.
LSTM-based system for ANWB to predict traffic incident severity based on real-time weather data. Achieved 92% accuracy with Streamlit dashboard.
Complete NLP system for Content Intelligence Agency: audio transcription with Whisper, emotion detection with RobBERT, and neural machine translation. 85% F1-score on Dutch content.
Computer vision system with MobileNet transfer learning for classification of 7 Dutch bird species. 95% accuracy with Figma prototype and A/B testing.
Machine learning system for market value prediction of 14,445 professional football players across 41 leagues. RandomForest model with position-specific feature engineering achieved 76% accuracy.