Washington, District of Columbia, United States
• Matured NLP products focusing on name entity recognition (NER), transformers, and sentiment analysis leading to an increase of product sales by $3 million.
• Built end-to-end machine learning pipelines, utilizing Scikit-learn for supervised machine learning (Naïve Bayes, SVM) and TensorFlow for deep learning (LSTM, RNN) models, used for text classification resulting in an increase of analyst efficiency by 33% due to automation of tasks.
• Tested software for bugs and operating speed while documenting the process and logic of changes which helped reduce total run time of our programs by 23%.