Apr 2017 – Oct 2017 · 6 mos · Hyderabad Area, India

Aug 2015 – Mar 2017 · 1 yr 7 mos · Hyderabad Area, India

Sep 2014 – Jul 2015 · 10 mos · Gurgaon, India

• Early engineer for a startup that got acquired by Snapdeal. I'm responsible for adding features to the core recommendation as a service platform that generates real time product recommendations which are being used by e-commerce sites with large customer base in India.
• Worked on a targeted search platform that generates users/cohorts recommendations for complex queries on user behavioral attributes or targeted catalog keywords. Implemented map reduce jobs to generate user’s behavioral attributes and predictive interests.
• Developed a scalable system for generating personalized best sellers / trending products and new arrivals based on user behavioral data. This system can personalize best selling products across combinations of multiple dimensions like category, store, gender etc. Experimented with several score decay functions to filter out old/irrelevant catalog items for each user.
• Technologies: Apache Spark, Hadoop, Elastic Search, AWS, Memcached, Java, Perl, Shell scripting

May 2013 – Jul 2013 · 2 mos · Bangalore

• Built a classification model to detect adult and fraud apps in Amazon's Appstore. Performed feature engineering, model design and parameter tuning using NLP techniques. Final model improved the overall accuracy of the adult and fraud apps flagging system by 30% on internal datasets.
• Designed an e-commerce website multi-classification model for seller services and wrote several automated scripts for website data feature analysis and model tuning.
• Technologies : Python, Java, Lucene, Shell scripting
• Team : Amazon Machine Learning, Platform Development