Course: Unsupervised Machine Learning & Data Mining under Professor Roi Yehoshua

Modelling Framework Optimisation: A product based customised set of ~260 models built in R using H2O GLM, targeting top n% buyers with maximum propensity to purchase in subsequent quarters.Outcome: ~7% uplift in top target bucket. Reduced Overfitting by incorporating distribution of Variable Importance in model evaluation.Integration of new algorithm into existing framework as an additional feature.

Course: Unsupervised Machine Learning & Data Mining under Professor Tina Eliassi-Rad.

Developed algorithms for video slot games, Max Power Series- Dragon Edition (Product Madness), Parisian Fortunes (Slotsburg), Ringo’s Riches (Big Fish Casinos), Collaborated with Developers, Animators and Testers to take the game into the production.

Developed Weight Prediction Engine, a framework to predict weight of e-com packages via consuming the historical data of clients, product features etc. Model deployed with 95% coverage and 90% accuracy.Applications: T+1 Provisional billing, real time cost realization tool, Weight Capture compliance check via predicting large shipments, Identification of weight capture anomalies.