Mar 2018 – Mar 2019 · 1 yr

• DNN inferencing platform using FPGAs.

Apr 2016 – Feb 2018 · 1 yr 10 mos

• As a part of Mailbox intelligence team, worked on FocusedInbox service for Outlook. Focused Inbox classifies the emails based on customer triage behavior and puts important emails in the Focused view and rest of them in Other view. This service is currently used
• by hundreds of millions of customers helping them save few hours of productive time every week.
• Contributed to various parts of the service like classification, training, model improvement, A/B testing etc. while keeping the latencies within few milliseconds.
• Worked on Online and Offline experimentation of various models and features.
• Migrated the existing live FocusedInbox service to a more agile platform for faster iteration.

Sep 2015 – Mar 2016 · 6 mos

• Azure Data Factory

Feb 2014 – Aug 2015 · 1 yr 6 mos

• Azure Data Factory

May 2013 – Aug 2013 · 3 mos · Palo Alto, California

Jan 2013 – Dec 2013 · 11 mos · Stony Brook University, NY

• Currently working as a Research Assistant under the advisement of Prof. Steven Skiena on BigData Analytics.

Jun 2008 – Jul 2012 · 4 yrs 1 mo

• As a part of the Autosupport Data Processing Team, I worked on design and Development of a highly scalable solution for ASUP data processing infrastructure using various big data technologies/frameworks (Hadoop, HBase, MapReduce, Solr, flume etc).
• Involved in the design and development of a MapReduce based framework for data extraction and loading.
• Involved in HBase schema design and development of data collection and ingestion module using flume.
• Involved in performance tuning of jobs on Hadoop cluster.
• Developed state machine based text parsers using Perl for data extraction from Autosupport messages received in order millions.

May 2007 – Jul 2007 · 2 mos

May 2006 – Jul 2006 · 2 mos · IIT - Kanpur

• Studied and implemented various image processing algorithms for ‘Image Apprentice’ – a product of Advanced Digital Imaging Solutions Laboratory (ADISL), IIT-K.