• Design and build TimeSequence Builder Library to capture and emit metric latencies (UPML) in near-real time among Alexa microservices, using Apache Flink, Kinesis Streams, Kinesis Datanalytics Analytics and Java.

• Implemented logic for generating UPML metric latencies, and built pipelines for internal teams to consume and dashboard the metrics, using Amazon S3, PySpark and Cradle.

• Worked on optimizing and stabilizing Apache Flink application, using AWS CodeGuru profiler. Resolving the serializations issues to ensure high availability of the system.

• Worked on implementation of data pipelines to query UPML metrics from datalake (S3), using AWS Lambda and Amazon Athena, in Java.

• Automated pipelines and Infrastructure stack creation for CI/CD using AWS CDK, in Typescript.

• Implemented End to End testing for Alexa Music microservice using Cucumber JVM, Gherkin and Java.

• Involved in automating data pipelines using Apache Airflow and Python.

• Experience in handling Alexa Music BI systems, as a single technical resource.

• Worked on resolving day to day support issues, migrations, and managing pipelines, in a fast paced software environment.

• Developed Metric Anomaly Detection in Python to forecast customer and dialog metrics for Alexa Music OP2, using SARIMAX model.

• Involved in engineering and modeling Alexa Entertainment BI systems, including Redshift cluster, security groups, users and user groups to better manage the costs between Alexa Teams. Migrated 150+ tables (over 150TB in storage) and ETL’S to new database using Python and Datanet – a Big Data internal tool for building pipelines in Amazon.

• Implemented Redshift Spectrum and data Retention policies by using Time Series Tables to resolve diskspace and performance issues in Redshift cluster.

• Integrated various Amazon services such as Hoot, Bumblebee, CloudWatch, Firehose stream, Redash to work seamlessly with new Database.