Vibhu Kundeti

• Evaluated AWS service setup options and setup infrastructure for a Fargate service over ECS to provide ad load recommendations for Prime Video titles based on customer/title metadata
• Supported internationalization effort by creating and testing ETL pipelines to launch title recommendation
  models in UK/Germany
• Created AWS resources and permissions, developed code changes, and setup integration and load tests to support launching back-end service in Japan
• Developed an ETL pipeline along with changes to existing featurization jobs to support the launch of a new title recommendation model architecture
• Exploring/experimenting with the impact of ads being introduced to Prime Video on the user experience

• Collaborated with ML team to design, prototype, and evaluate a title recommendation model for new users to
significantly drive engagement on the Freevee App
• Developed ETL jobs in an Apache Airflow pipeline to generate and maintain the model using incoming and
existing customer playback data with AWS Tools (Glue, EMR, S3) and Python
• Designed and developed infrastructure changes for a Java back-end service to support title recommendations
from multiple models and redirect new customer requests to the new user model alongside JUnit tests and

• Automated common client workflows using Javascript and the MicroStrategy API involving sample API requests,
  informative test cases, and documentation
• Collaborated with the UX team to integrate the feature within an internal developer site