Sandeep Thonukunuri

Projects:

Enterprise Claims Streaming Application:

· Developer multiple data ingestion batch jobs between different sources systems to downstream databases.

·· Building the JSON structured data objects and loading them into PostgreSQL and Mongo DB.

. Heavily used case classes, spark Actions/transformations, dataframe joins, window functions.

.Kafka Streaming sessions to connect topic and to publish messages for real time application.

· Monitoring the Kafka Streaming jobs and provide real time updates to the business teams.

Ø CIP Batch Extracts: (Legacy Mainframe to Spark/Scala Reporting)

· Using Bean IO framework, generated Fixed length and Pipe delimited files as per the vendor specification.

· Used Spark 2.0 API to integrate with Bean IO.

· Configured XML mapping files and Bean classes in order generate the Fixed length files.

· Used Spark-Streaming APIs to perform necessary transformations and actions on the fly for building the common data model which gets the data from HDFS.

Ø Reconciliation between HBase and Mongo

· Identified rowkey and column family of the Source HBase table.

· Define a catalog schema for the HBase tables.

· Used mongo spark connector to connect Mongo database.

· Created a CSV file using CSVWriter.

· Listed out the data discrepancies and Prepared a excel sheet with missing data in Mongo

Ø ETL Generic Framework:

. It's a configuration based framework where developer doesn't required to code for straight data ingestion.

Ø UI for the Florida Blue Login Page:

· Worked on changing the UI components for the Florida Blue website Login page this will be shared across the organization for Members, BA’s and Agent.

· Using Servlet and JSP added some UI components such as, forgot user ID, Forgot Password, Notification Page after resetting the password, Need Help and Accessibility functionalities.

Business Continuity- Disaster Recovery

RxSS business model is heavily reliant on client providing critical files Formulary, Network, Plan for accurate pricing model for the Members. A process has been developed to relay 100% only on internal claims and analyze the data in a bigger picture and provided a pricing model.

· Created multiple DeltaLake table using Scala/Pyspark.

· Devolped databricks Notebooks using multiple commands and created a Notebook sequencer to execute the multiple notebooks.

· Performance tuning of Spark Applications for setting right Batch Interval time, correct level of Parallelism and memory tuning.