Jilin He

• Designed and implemented 4 Python services and packages using PHP, JavaScript, and SQL to automate lifecycle management of smart data models by 50%, reducing processing errors by 20%. • Engaged in daily maintenance and FIWARE Open Source Community activities, including troubleshooting, issue resolution, and contributing code to core components.

• Developed a data processing pipeline to transform web-scraped paragraphs into contextual text blocks, employing segmentation, tagging, and annotation. • Created novel keyword extraction datasets for non-technical articles, comprising 10,000 contextual text blocks sourced from companies’ descriptions and 50,000 blocks extracted from their annual reports. • Achieved a keyword labeling F1 score of 94% via extensive model training with CNN, CRF, BiLSTM, BERT, and RoBERTa.

GPA: 1,3/sehr gut - Relevant Courses: Database Technology, Big Data Systems Project, Neural Network Simulations for Linguistic Research, Machine Learning for Remote Sensing - Thesis topic: “Towards a Novel Classification of Table Types in Scholarly Publications”

GPA: 16.41/20.00 • Relevant Courses: Applied Statistics, Mathematics for Data Science, Relational Databases, Machine Learning Algorithms, Deep Learning, Hands-on Natural Language Processing, Innovation and entrepreneurship (I&E) • Award: EIT Digital Master School Scholarship