https://www.oracle.com/integration/data-integration

https://www.oracle.com/big-data/data-catalog/

https://www.oracle.com/big-data/data-flow/

 Spearheaded the development of a highly scalable feature store to manage the creation, storage, and serving of features, ensuring efficient data processing for both offline model training and real-time online predictions. Implemented on-demand Milvus, an open-source vector database, to store and index embedding vectors for high-performance similarity search. This integration enabled personalized item recommendations based on user queries, significantly enhancing the recommendation system's overall performance and scalability

𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Designed and implemented a high-performance platform handling over 3 billion API requests daily. It delivers real-time insights with a 95th percentile execution time of 3 milliseconds, enabling rapid iteration and optimization. Features include user segment and cluster creation for targeted experimentation, enhancing performance across diverse user groups. This platform fosters collaboration with transparent, actionable data. Contributed to DRUID by adding a parquet extension, a P-Value aggregator, and fixing bugs.

𝗚𝗶𝗺𝗲𝗹 𝗗𝗮𝘁𝗮 𝗔𝗣𝗜: As a core engineer, developed and deployed GIMEL, an Apache Spark-based big data framework. GIMEL provides an efficient abstraction layer for data access across storage systems and integrates with Jupyter notebooks for interactive exploration. My contributions, including the GIMEL Thrift server, enable robust, secure data access and streamline the data science workflow through rapid prototyping and visualization.

𝗨𝗻𝗶𝗳𝗶𝗲𝗱 𝗗𝗮𝘁𝗮 𝗖𝗮𝘁𝗮𝗹𝗼𝗴: Developed PayPal’s Unified Data Catalog, a metadata system for governing data and AI assets across files, tables, models, and dashboards in both on-premises and cloud environments. As a founding engineer, I was crucial in designing and implementing this comprehensive system.

𝗡𝗲𝘅𝘁𝗚𝗲𝗻 𝗗𝗮𝘁𝗮 𝗠𝗼𝘃𝗲𝗺𝗲𝗻𝘁 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺: Designed a scalable, autonomous data movement platform with schema registry and automated TSDB monitoring. Technologies used include Apache Gobblin, Airflow, Aerospike, Oracle, and Spark.

𝗙𝗶𝗿𝘀𝘁 𝗣𝗮𝗿𝘁𝘆 𝗧𝗿𝗮𝗰𝗸𝗶𝗻𝗴 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 (𝗙𝗣𝗧𝗜): Led the development of an internal client and server-side event tracking platform. This platform includes APIs and libraries for efficient data capture, processing, and real-time report generation. It is essential for tracking web traffic and KPIs for various PayPal products and marketing campaigns.

Involved in end to end development of a Data Warehousing product from the scratch for a leading DW company