Machine Learning & Nlp Intern
- Worked on a joint project between Columbia University & Muck Rack for press-release detection- Generated features for 350,000+ press-release data samples via token lemmatization on article plain text, employing sentiment and TF-IDF analyses- Achieved 97% accuracy after training and evaluating Logistic Regression, Decision Tree, and XGBoost machine learning models- Applied text embeddings via Sentence Transformer and K-means clustering for article plain text, reducing runtime by 60%