Modelled Extractive based Text Summarization Product for client using TextRank Algorithm and LSTM model using GloVe and Google embeddings.

Built a Chatbot using Google’s Dialogflow for Broadridge, capable of answering finance related questions. It answers generic questions as well. Now building a chatbot model from scratch for finance specific corpus using Weighted Word Embeddings with TF-IDF.

Applied SVM and Logistic Regression for Employee Attrition Prediction, gaining 0.72-0.81 f1-score, later taken over using Random Forest Classifier producing results with 0.92 f1-score.

Utilizing sentiments of Stock Record News from the news aggregator was important for assisting the prediction of stock record growth. The news sentiments were captured using NLP’s VADER. The code developed was later used as a ‘library’ for Stock Record Growth prediction along with XGBoost model mentioned above.