• Developed a context-based recommendation system for Evernote notes using various NLP techniques (LDA, TF-IDF, Word2Vec, FastText).
• Implemented various document classification algorithms using Spark with a top prediction accuracy of 90%
• Created machine learning pipelines using Apache Airflow, SQL, and Google Cloud Platform services.
• Optimized Churn and Retention models by applying feature engineering and hyper-parameter tuning using Scikit-learn, Tensorflow, and ML-Engine, increasing prediction accuracies by 8-10%
• Lead the development of a cryptocurrency payment portal for Evernote’s premium service.
