Mar 2021 – Aug 2021 · 5 mos · Bangalore Urban, Karnataka, India

• Working on machine learning models to solve problems in multimodal content understanding and search domains.

Jun 2019 – Mar 2021 · 1 yr 9 mos · India

• Working on machine learning models to solve problems in Ads Targeting, Ad Relevance in Walmart E-commerce.

Jan 2016 – Jun 2019 · 3 yrs 5 mos · Hyderabad Area, India

• Worked on designing and developing data science solutions to solve business problems in the fields of NLP, Text Mining, Predictive Analytics.

Mar 2015 – Jan 2016 · 10 mos · Hyderabad Area, India

• Worked on building end to end machine learning pipelines, starting from Data Analysis, building POC models, publishing performance metrics, deploying scalable solutions to production to solve business problems.

Mar 2014 – Mar 2015 · 1 yr

• Work involves design, development of content-based recommendations for Music Kramer. It involves recommending similar music artists, albums, songs for a given artist, album, song respectively.

May 2012 – Mar 2014 · 1 yr 10 mos · Bangalore

• In Veveo, I broadly work on named entity reconition, text categorization, building knowledge graph, phrase extraction, Mining unstructured data from web, deduplication of named entities, entity pruning based on global and location relavance features, query normalization, data normalization, determining quality of lexical and phonetic variants of phrases using knowledge graph.
• Few of the modules that i worked in veveo are:
• 1) Kramer Entity Recognizer: Involves building scalable data structures for storage and efficient retrieval of large number of entities, support for federated ER, pruning entities based on relations between entities in current query context , query normalization, query stemming, deduplicating entities, entity pruning, phrase extraction from the query etc.
• 2) Building Knowledge Graph of entertainment domain : It is a graphical representation of entertainment domain data (which includes movies, tvseries, crews, roles, sports teams, tournaments, players etc.,) which elicits relationship among entities. Some of the challenges that i worked on are enriching meta information of entities from unstructured data using NLP based approaches, merging entities from various sources, global relavance score calculation etc.
• 3) Generic Folding: It is a text categorization technique. For a given record , it associates most relavant entities.

Jul 2010 – May 2012 · 1 yr 10 mos · Pune Area, India

• I am part of the Veritas Dynamic Multi-Pathing (VxDMP) product development team.
• Development in C