Member of Hadoop and NoSql team.
Optimizing end-to-end data transfer from the database cluster to the Cohesity distributed file system via the Avro IPC server, Hadoop backup/restore adapter, and Grpc API server.
Leading and designing Hive view protection involves building dependency graphs between views and tables, leveraging the Calcite parser for Hive SQL queries, and employing multi-threading for concurrent restore of tables and views in their dependency order.
Reducing the data source registration time for one million Hive tables from 300 to 20 minutes by leveraging batch processing, the producer-consumer multi-threading, and shorten data conversion.