• Conducted a comparative study of traditional recommendation systems and their neural network variants. • Implemented Content based, Collaborative Filtering (user-user, item-item), Baseline model, SVD, Neural Collaborative Filtering, Variational Autoencoders and k-means using PySpark on the 25 million movie lens dataset for recommendation purposes. • Observed that a combination of linearity of Matrix Factorization and non-linearity of deep neural nets results in better modelling of user-item interactions.