Umesh kumar

Scaled up a legacy WALLET monolith system of 29TB SQL based service into microservices to support 10X scale by breaking into microDBs and companion microservices without compromising on transactional guarantees. Re-architectured ATS(Logistics) service to be more cost efficient and reduced latency by 35% by identifying the best ROI based approach. Achieved through SQL queries. This effort resulted in cost saving by a factor of 20%. Has been actively working on GenAI based initiatives - Helped team to adopt to github copilot and chatGPT in our day to day processes to be more productive. Developed a POC on auto code gen and code review module.

Designed/Built bottom up framework for help systems and communications, customer engagement platform. Designed/Built MASL (mobile aggregation service layer) for our mobile apps which enabled mobile app developers to bring up scalable aggregation nodes instantaneously.

Our team mainly focuses on backup solutions for any type of workload and am part of backing up on-premise file/folders to Azure. As part of the day-to-day activity in this team, I concentrate on reducing the costing we incur to storage as part of storing large amounts of data (as large as 950 PB) through optimizations of storage format, better GC etc. Redesigned the existing storage format of customer data which saved us more than 75% of storage costs. Leading a team of three engineers in getting this solution implemented. Platformizing the product designed, developed v2 of data protection platform which would enable lot of workloads seamlessly onboard on to backup product.

With current market trend, Adobe decided to invest in mobile space by providing customers mobile companion applications with slimmed down feature set to that of desktop application. Designed a composite format with which applications can download and operate on part of file. Demonstrated the design and implementation of SSO(single sign on) across all Android Adobe apps installed on a single device using AccountManager. Native Extensions support for FL is a programming framework that enables Actionscript code running in AVM1 or AVM2 to call and be called by native applications and libraries written in languages such as C/C++. This enables content developers to handle situations when an application cannot be written entirely in Actionscript. Prototyped ideas related to improving rendering performance, optimizing and enhancing the native extensions framework, some of them have been productized.

Worked on design and implementation of PL/SQL procedures for identification of inconsistencies in the INSEAD Business Data and for correcting those inconsistencies.