Sunnyvale, California, United States
Linux Kernel Networking Prev: Google Kubernetes Engine (GKE)
Working in the Distributed Systems Research Group on YScope Compressed Log Processor (CLP). Currently working on designing a distributed log compression architechture to process logs from all customers using CLP Cloud
Developed a service to deduplicate AWS bill aggregation triggers to reduce a growing problem of redundant bill aggregations.
My role centered on R&D for a complex application which acts as a data pipeline to aggregate and process account level messages for customers, to presented in the mobile app. Developed and compared two architectures, server backed and serverless on AWS: Serverless stack processes from a Kafka topic via a scheduled AWS Lambda and stores them to DynamoDB (NoSQL) table. A separate Lambda Function + API Gateway app is used to retrieve messages from the database. Server-backed stack processed messages via a concurrent Java microservice that dynamically integrates new message sources (Kafka topics, Realtime APIs, etc.) by picking up endpoints in application configuration. Gained a good understanding of Infrastructure as a Code (IaaC) tools and practices, Jenkins pipeline creation and optimization, AWS Lambda, AWS DynamoDB.
Integrated a machine learning model to the application fraud decisioning process by adding decisioning layers into a Scala + Play microservice. Added capability to flow decisioning data to a kafka topic and a data lake for model validation and analytics (via Snowflake). Saved 10% of the total time taken to decision fraud on a credit card application by redeveloping an old Scala-based microservice in Java and Spring Boot. Refactored and maintained Jenkins pipeline which used enterprise plugins to deploy application to AWS.