• As part of Consumer Platform, develop scalable data integration/orchestration platform to simplify development and fast shipping of the consumer features.

• Designing and developing highly scalable systems for PayPal P2P flow using open source and PayPal's technologies.

• Built highly scalable generic data platform to host insights pertaining to the real estate listings/homes. Platform supports low latency (under 15 ms) and highly available REST APIs supporting live Realtor.com traffic.

• Architected/developed the production pipeline for the Photo-First initiative for the Realtor.com. As part of this, images/photos pertaining to the home listings are classified/categorized under different categories like External, Bedroom, Bathroom, Living Room etc as displayed on Realtor.com. This feature has direct impact on user experience and thus CSAT score.

• Improved the accuracy of the MLS listings description based tags. These tags were based on the regular expressions. New system used NLP based ML model to improve the accuracy and thus improved overall user experience.

• Hero image selection for the certain listings for which images/photos uploaded were not visually nice. The system using Click Through Rate (CTR Score) of the images would identify the top two images to be displayed as Hero Images on SRP (Search Results Page) and LDP (Listing Detail Page).

• Deployed a pipeline to extract the short key phrases/features from the listing description using NLP to present the key features of the listing/home to the user. This module was named "Homes with Special Features".

• Designing and developing a highly available, distributed shared data service platform. Data is exposed through REST compliant services. Platform supports a flexible query DSL to perform the searches on the data across highly utilized applications.

• Couchbase is being used as data-store. Couchbase querying is done using N1QL and 'get by keys'. Java Couchbase SDK is used as an interface to Couchbase.

• Data Change Notification mechanism let's consuming applications to get notified of the data change in the system. Apache Kafka us being used as a backbone for Data Change Notifications.

Solr is used as the search platform to perform basic/complex searches.

Worked on a tool to monitor the progress of Walmart’s Item ingestion pipeline. Item ingestion pipeline comprises of multiple resource tiers like Item Pricing, Ship Price, Ship Methods, Item Store. Objective of the system was to monitor the changes made to the items attributes and making sure the changes gets reflected in appropriate target system(s). In case of failures, item has to be re-ingested. It also supports various analytics reports presenting clear picture of the ingestions done through the pipeline.