• To develop a Recommender System that suggests groceries based on customers preferences & their purchase history
• Performed memory optimization & EDA to identify correlations and anomalies that aided in model development
• Implemented K-Means Clustering (based on Elbow Method) for personalized recommendations to similar users
• Generated recommendations leveraging association rules derived from lift score for every product pair in clusters
• Constructed an NLP Search Engine for products based on cosine similarity between count vectors post stemming
• Built SVD model for new users | Reduced RMSE by 60% | Used diversity metric for overall varied recommendations
• Deployed Flask based Web app providing recommendations to the users based on their ratings and past purchases