For 3 leading Telecom customers:Upsell propensity model - Plan recommendation>     Developed and deployed a machine learning model to help a telecom provider target customers to up-sell based on the propensity of the customer to respond to the proposed campaign>     Used the XGBoost algorithm to predict the propensity scores for the plans>     Deployed the system in the AWS Cloud environment and made a real-time prediction system from the model in AWS SageMaker, with sub-second latencyCustomer Lifetime Value Prediction>     Developed and deployed a machine learning model to help predict the LTV of customers and segment them for loyalty programs and other marketing strategies >     Used multiple probabilistic and ML models to predict the CLTVFor a global CPG manufacturer -Assortment Optimization> Built a demand transference model and rationalization model (to score complexity in the product assortment in terms of Sales, Profitability, Supply chain complexity, Incrementality) to measure the possible impact of rationalization, thereby help ease decision-making for adjusting the assortmentProduct Innovation - FP&A> Orchestrated the development of a self-serve FP&A WebApp product that transformed the product innovation process, enabling Financial Strategists to estimate P&L numbers for innovations and renovations proactively, before they hit the market. > Developed modules for sensitivity analysis, cashflow simulation to enable profitable, strategic, and faster product innovation for multiple functional groupsSearch Relevance Prediction> Redesigned the process for evaluating the performance of customer's online search algorithm by creating a model that can be used to measure the relevance of search results> The solution was developed with advanced feature engineering and model ensembling to enable a highly accurate prediction of relevancy score, thereby minimizing the manual effort required in search relevance evaluation.