Prakhar Raj
About Prakhar Raj
> Data Science enthusiast with an experience of 2+ year working in the Fintech industry. Currently working in a product based company to primarily improve Order to Cash process in the B2B scenario by automation, prediction and gathering insights using Machine Learning and Deep Learning. > Contributed to building and scaling a Global B2B Payments and Receivables Network to enable digital collaboration between suppliers and buyers by solving the problem of Customer Linkage or Entity Management to uniquely identify customers across all accounts using extension of Hierarchical Clustering. > Built Deduction Validity Anomaly Detection engine using ensemble of bagging/boosting tree-based ML algorithms with an invalid capture rate of more than 80% which is consumed by several Fortune 500 clients since it reduces their manual effort by significant amount and thus improving Deduction Management System. > Worked on self-sufficient and smart system to perform Dynamic Segmentation of customers by analysing their past payment behaviour and revenue from customer using Unsupervised Learning methods. > Built ever-learning and intelligent system for generating real-time prioritized work-list of customers using Supervised and Unsupervised Learning techniques thus improving the Collection Management System of clients. > Experienced in Data Visualization, Data Analysis, Predictive Modelling, Forecasting, BigData with PySpark, NLP problems, Deep Learning and cloud services like AWS for deployment of solutions. Long story short, I find solutions.