Mar 2021 – Apr 2023 · 2 yrs 1 mo · Bangalore Urban, Karnataka, India

Nov 2016 – Feb 2021 · 4 yrs 3 mos · Bengaluru, Karnataka, India

Jun 2015 – Oct 2016 · 1 yr 4 mos

• Working on a product which is a dynamic data warehouse on hadoop, which features data augmentation to multi dimensional cubing
• Technologies : Java, Hadoop, Hive, Apache Kylin, Hbase, R
• Building an intelligent data transformation engine with interactive feedback in visualize as you build flow. Responsible for the backend of the engine from design to deployment.
• Leading a team responsible for all aspects of the data augmentation and cube building.
• Modified Apache Kylin to support snow flake schema, implemented a solution to support high cardinality dimensions
• Integrated Kerberos, Apache Ranger and enabled secure impersonation for our product
• Redesigned the data augmentation engine for it to be simpler, configurable and scalable.
• Dependency based parallel joins and transformations.
• Helped simplify the product installation for easier adoption of the tool
• Worked on a library on R to expore our data

Apr 2012 – May 2015 · 3 yrs 1 mo

• Conquest (September ‘13 – May ‘15)
• Technologies: Java, Hive, HBase, Hibernate, MySQL, Fusion Charts, Phoenix
• Conquest is an in house developed OLAP engine. It has two main components, a data transformation layer to create tenant specific facts and dimensions, and a reports execution engine which is used to get data out of the warehouse
• Worked on a new data transformation pipeline using the big data stack involving hive, hbase and phoenix
• Worked on an active passive switch over mechanism for higher availability of the system
• Dashboards working off summary tables for really quick response times
• Parallel query execution based on the query dependency for UI responsiveness
• Partitioned the data sync files based on tenant geo zones hierarchy. This reduced billing costs for the company by
• cutting down on network transfer charges in AWS and also improved experience on our client module.
• EMF and Points Engine (Nov ‘11 – July ’13)
• Technologies: Java, Spring, Hibernate, MySQL, EHCache, JUnit
• EMF is the core multi tenant OLTP module for handling all business events, with ability to perform real time incentivisation and communication.
• In house rule engine developed with a custom expression language, with tenant specific grammar
• Responsible for the loyalty module which is at the core of the CRM for capillary which handles credit-debit of customers
• Offline batch processing module for points expiry, tier downgrades, reminders, scheduled offers etc
• Migrated the online legacy OLTP credit-debit system, in phases, onto this module seamlessly with configuration and data sanity and almost no down time
• Migrated an existing Dynamic Coupon System based on Drools

Jul 2009 – Apr 2012 · 2 yrs 9 mos

• InTouch Client (Sep ‘10 – Nov ’11)
• Technologies: C#, MS SqlCe, Java
• The client is responsible for collecting billing, customer, redemption etc information from the POS and syncs it with the InTouch servers over HTTP apis
• Re-factored the store server module to support multiple internal client terminals and a shared database
• Developed the background sync module to improve responsiveness and availability of the client
• Developed a deployment server which helped ease the binary releases and save lot of time
• Involved in numerous features, improvements, caching, bug fixes and UI/code refactoring, database abstraction for working with sqlce, ms sql server and sql lite, internal cron for jobs, async logger with event ordering.
• Intouch Server (July ‘09 – Aug ’10)
• Technologies: PHP, MySQL, Thrift, Apache, EC2, Mondrian
• The server is responsible of receiving data from the client. Modules for Administration, Campaigns, Loyalty, Inventory, Rule Management, Reports, Integrate with Communication engine for sending SMS to customers.
• Implemented caching of configurations and frequent objects to reduce load on database to a huge extent.
• Reduced database load by query optimizations, index creations / modifications
• Developed catalog sub-module, fraud detection and access control. Involved in numerous features, improvements, bug fixes and UI/code refactoring. Put in place a performance parameters monitoring and recording, log file parser for to extract information on a per request thread basis

Jan 2009 – Jul 2009 · 6 mos

• Worked on a Topic based 'Expert' extraction mechanism based on Email communications within an organization. Analyses the email threads along with the attachments contents to associate 'expertise' to an employee. Given a query, the module was expected to show up people who can be contacted. Technologies : Perl, Java, Lucene, JUNG (graphing) etc