Worked on CPEntity which is a control plane storage library used by Tupperware(Kubernates equivalent)
Developed APIs to integrate with a low-dependency multi-shard database(Delos), supporting services such as History, Indexing, and Streaming.
Designed and implemented an asynchronous subscription library, enabling users to create synchronous streams and subscribe to filtered entities. The library efficiently multiplexed a single data stream across multiple users.
Refactored the History service to leverage the new subscription library, improving performance and maintainability.
Created a live consistency checker for the History service which guaranteed there was no corruption of data in the pipeline from the main database to History.
Built an end-to-end metadata system tracking every version of entities, allowing users to associate metadata with changes and retrieve historical hierarchies through the History service.
Improved reliability through Oncall support of CPEntity
Software Intern
Equinix
05.2021 - 08.2021
Designed and implemented a process in which microservices can separate their transaction flow into multiple stages and send information about each to a separate microservice.
Deployed a microservice using the Spring Framework that reads status events from other deployed microservices through Kafka and stores them in a Postgres database.
Created an API to access stored status events, group them using different criteria including transactional IDs, and append contextual data to each event.
Created a standalone status event producer that can be easily added to any microservice and provides capability to break down a flow into a set of stages that each emit a status event through Kafka.
Software Intern
Avi Networks
06.2018 - 08.2018
Created an automated process to extract virtual machine statistics from customer tech support packages to identify I/O spikes and track usage of resources.
Extended an existing Postgres database to store different statistics about the virtual machines across time to assist in identifying trends about I/O operations and resource usage.
Worked with Apache Superset to create visualizations of different I/O operations like reads and writes across time as well as visualizations of resources such as CPU usage across time.
Identified specific jobs and job types that were inefficient in terms of I/O operations or exceeding resource usage on the virtual machines.