Shubham Sanghi

Studied research papers, designing and evaluating multiple algorithms and outlier detection methods, culminating in the creation of a robust fraud detection model and collusion detection algorithm, effectively detecting with a recall of 79% addressing fraudulent players within the game. Implemented a breakthrough discovery of the 'aha moment' in the game, resulting in a substantial 23% increase in long-term consumer retention at the 30-day milestone. Conducted detailed analysis of user purchasing behavior in physical retail settings, strategically optimizing SKU placement to achieve a 7 basis point improvement in upselling performance. Created and validated dashboards to assess the impact of A/B experiments across various metrics, meticulously collecting and analyzing data according to stakeholder requirements, and identifying critical customer parameters such as resilience, sensitivity, retention, lifetime value (LTV), and engagement within distinct cohorts. Provided actionable recommendations based on the findings. Developed a system to enhance the migration of high-value users from low to high revenue applications, successfully deployed and yielded a 5-10% increase in revenue and daily active users, while simultaneously reducing the cost of acquiring new users.

Project Lead in predicting the next character from the previous set of words or sentences with the help of convolutional neural network, LSTM and adam optimizer. (Optical Character Recognition) with a precision of 89%.