Developed a service in C++ to reduce user’s ODS usage by automating usage analysis & producing efficiency opportunities for users

Service will bring down the ODS’s YoY growth by 10%, translating into savings of around 25% of active time series and 700 hosts next year for in-memory storage.

Worked on a distributed design for the service to run simultaneously on more than 50 hosts using Thrift and Chronos, each processing millions of rows of data and analyzing & generating opportunities for users.

Onboarded the service to the central demand management system to enable teams to use resources efficiently.