– Developed an automated pipeline using AWS Lambda, Kinesis, Dynamo DB, Neptune Database, and other AWS tools to ingest sports datafrom 3P Data Providers. Implemented a 24/7 polling mechanism calling APIs to ingest real-time data during sport events.– Designed multiple layers in the infrastructure for transforming raw payload into internal sports data models by key-mapping external IDs to internal IDs. Created a new technical approach called doppelgänger for finding duplicate internal IDs and merging them from multiple providersbased on elastic-search.– Built a REST API upon Neptune Graph Database using Amazon API Gateway and AWS Lambda, enabling downstream teams to retrievetransformed data. Achieved 99.9% API availability and utilized AWS X-Ray for tracing and analyzing API calls.– Onboarded 4 new sports to Prime Video Sports in the period of 9 months by modeling data frameworks and setting up ingestion infrastructure. Reduced onboarding effort from 45 days to 5 days. Currently working towards creating an automated approach to further reduce onboardingeffort by saving 89% of the time required.

– Developed a web application using JavaScript and HTML/CSS to extract depth and spatial data from video datasets of Visible Light Communication Systems, and analyze raw data of transmitted and received bit-stream using Digital Image Processing module. Optimizedthe data capturing and analyzing process of the web application to reduce dependency time by 74%. (bit.ly/vlc-webtool)– Designed and implemented a Unified Data Access Layer using Hasura to create a master data source for all data at the Dice Lab, enabling faster analyses and decision-making by centralizing data sources and providing a single endpoint for querying. Leveraged the capabilities of GraphQL to easily connect multiple databases and pull resources from them, while Hasura’s automatic schema generation simplified APIcreation and implementation of authorization rules. (quip.com/OnFiAvDVUo3V)

– Developed android and iOS application utilizing Optical Character Recognition to identify Expiration dates from input images of everyday products to notify users before its expiration.– Integrated Google Calendar to sync dates and push event alerts on multiple devices. Optimized storing product details using Barcode Scanner in firebase to remove data redundancy. Leveraged knowledge in Flutter, Machine Learning Module in Firebase, Git. (bit.ly/expiry-date-scanner)