Machine Learning Intern
• Coordinated efforts to gather and categorize data from 150+ newspapers, ensuring the collection of diverse content types; this initiative resolved the dataset's richness and usability for longitudinal analysis.• Pioneered and optimized the BERT configuration parameters on the CNN-Daily Mail dataset, achieving a notable increase in performance metrics, making suitable for real-time news summarization applications.• Extracted and analyzed over 500 data features from complex datasets, leading to a 20% improvement in predictive model accuracy for NGO projects, enabling better resource allocation and project impact assessment.