Developed a Twitter Recommendation Engine using Natural Language Toolkit (NLTK), LSTM, and other NLP techniques like Tokenization, Lemmatization and Deep Learning layers: Dense, Dropout, embedding etc.

Credit Card Fraud Detection: Visualization of data and understanding Outlier Removal Trade-off, using Dimensionality Reduction and other measures for highly imbalanced dataset, created a model for same.

Hotel Review Analysis: Found correlation between reviewer's nationality and scores using data-driven approach, performed sentiment analysis, fit a regression model on reviews and score to see which words are more indicative to a higher/lower score, and simple recommendation engine to the guest who is fond of a special characteristic of the hotel.

OCR (Optical Character Recognition): To build model for Handwritten Text Recognition using CNN, RNN, OpenCV (cv2), NumPy, TensorFlow, CTC loss and decoding layer.