Lead role designing and implementing a Cloud Platform for Big Data Enrichment. Provided and maintained predictive solutions for B2C companies enabling business users to accurately identify new market leads with detailed profiles and precisely score clients.
Sole contributor, developing native core engines for modeling and scoring. The modeling engine automatically builds predictive statistics models in supervised learning with any large volume commercial data – up to 60 million records with up to 700 variables.
Select Accomplishments:
Revamped old consulting predictive model in order to offer a cloud application solution for the current market, transitioning to the self-service predictive model.
Developed native core predictive modeling engine, including multiple algorithms such as classification, linear/nonlinear regression and Monte Carlo methods, etc.; high-performance core engine written with multithreading Window C++.
The improved customer wait time from one or two weeks to receive a model to a return of 30 minutes online model build.
Developed a back-end Java server to process large volume data for a cloud application, including data cleaning, matching, deduping, sampling, sorting, and query from Cassandra to prepare data for building look-alike and response models.
Utilized multithreading, Java data structure to achieve high performance, efficient memory usage for handling large-scale data of up to 130 million records; implemented on Spring Framework, JUnit, and Maven on AWS cloud platform.
Provided overall leadership to the entire software development product life cycle, including product specification definition; designing and implementing new features; GUI upgrades; managing the software release process; developing test plans, performing testing, preparing test reports; analyzing product performance and release scheduling requirements; and preparing appropriate PLC documentation.
