Feb 2021 – Mar 2025 · 4 yrs 1 mo

• Leading the data platform team at Eightfold to continuously scale, optimize and extend the data integrations to pull in the data at regular cadence from the external systems (like Applicant tracking systems, Learning management systems, Human Resources Information systems, Assessment systems, etc... - it's an ever-expanding list as Eightfold continues to scale).
• Eightfold has been building an AI-first approach to solve the problems in different product areas like Talent Acquisition, Talent Management, Talent Experience, Talent Analytics etc. And data platform team has been operating with the data mindset to continually build and improve data ingestion and data monitoring. Apart from that as we continue to onboard more customers, we build and enhance a data platform console to enable our solution architects to configure the platform to be tailored to meet the client's needs.

Apr 2020 – Feb 2021 · 10 mos

May 2018 – Apr 2020 · 1 yr 11 mos

Sep 2017 – Apr 2018 · 7 mos · San Francisco Bay Area

• *Sizmek acquired Rocketfuel in Sep 2017.

Dec 2015 – Sep 2017 · 1 yr 9 mos

• Leading the Modeling Infrastructure team to solve business problems by developing applications on top of existing data infrastructure (Hadoop, HBase, Hive, Storm, Kafka, Yarn).
• Architected and developed generic alert framework to help monitor 100+ data pipelines and in the process improving developer efficiency.
• Led initiatives to increase the robustness of bidding systems and avoid outages due to invalid data changes.
• Led the adoption of Apache Airflow within the team by debugging the operational issues and finding out the best practices to follow.
• Used scrum/agile methodologies to increase accountability and efficiency in the team.

Oct 2012 – Nov 2015 · 3 yrs 1 mo

• As part of Modeling Infrastructure team, I worked on several projects related to building data pipelines to support modeling research team in building ML models to predict the right ad for the user.
• Developed Real time Behavioral Targeting (BT) pipeline to append BT features directly from bid
• request to HBase data store and make them available for use by bidders within a second.
• Built Audience Guarantee targeting data pipeline using Hadoop MR to upload age-gender estimates to
• HBase store.
• Built feed processor to ingest client’s product feeds for serving dynamic creative ads.
• Built data streaming pipelines using Kafka/Storm for machine learning models used for ad predictions.
• Built framework to run scheduled queries and send automated email reports.
• Technologies used : JAVA, Scala, Spark, Hadoop, Hive, Hbase, MySQL, Vertica, Python, Kafka, Storm.

May 2011 – Aug 2012 · 1 yr 3 mos · Bengaluru, Karnataka, India

• I worked in Amazon Points team, part of Promotion Services Group.
• Worked on a project to design and develop a Points platform from scratch for Amazon to offer reward points to customers on all of its marketplaces. The platform would adapt to any marketplace, enable business to set up points-program in that marketplace where in shoppers would be able to earn redeemable points on the purchase of eligible products and provide scalable and low latency accounting service for shoppers. The platform is using DynamoDB, an AWS service storage solution, at back-end to ensure scalability of the Points Platform.
• Worked on designing and implementing a set of APIs exposed via Points Platform Accounts Service, based on Spring framework, to credit/debit points to customer accounts.
• Worked on designing and implementing a system to consume notification events continuously published by the Competitive Monitoring Team (CMT) to the Amazon Messaging Platform (AMP) host fleet which helps the Points Platform keep track of the points offered by the competitors on the same item. AMP is build on top of Apache ActiveMQ acting as the JMS provider.
• Improved the efficiency of entire team in resolving customer-account points related issues raised by the customers by investigating the general root causes for different types of issues and documenting the steps to follow to resolve those issues.
• Ownership of Points Offers Service, which is related to calculating the number of points to offer given an item or given a behavior.

Jul 2010 – May 2011 · 10 mos · Bengaluru, Karnataka, India

• I worked in the Content Solutions team, part of the Cloud Platform Group.
• Worked on a project of writing a data transformation plugin to convert data of EU Sports and EU Weather from the native XML format to Content Model format objects which were compatible to be stored into Hadoop grid and subsequently, available for offline quick data enrichment.
• Worked on a project of refactoring the Maven configurations of PEBL(Pacman Enabled Business Logic) packages, which were inter-dependent so that they can build into independent deployable jars.
• Owned the maintenance and fixing bugs of PEBL packages which were used as the processing units in a workflow defined to process the feeds coming in from different sources like ANI, Reuters etc.

May 2009 – Jul 2009 · 2 mos · Lausanne Area, Switzerland

• * Developed an interface for speech recognition using sphinx3 library that was later used by other members of CRAFT lab under the guidance of Dr. Frederich Kaplan and Prof. Pierre Dillenbourg.