I worked on the Google Cloud Platforms team on a Hard Disk Drive Management project that delivers core infrastructure for device firmware management, with a focus on reusability, testability, and observability.

I implemented a generic firmware version graph representation along with a core library that executes optimal traversals through that graph to a target firmware version.

I also Implemented a management library for ATA drives using a specialization of the generic framework

Finally, I created a prototype tool using all components with observability metrics visualized in a dashboard.

I trained under the performing arts company Academy of Villains during the entire summer of 2019. I was not only physically and mentally trained as a dancer, but also as a performer and an artist.

I worked with the Shopping Personalization team during this internship.

I developed a TensorFlow pipeline that creates a model for extracting product information from URL data in order to improve maintainability on product data, reliability of product information extraction, and efficiency of data storage.

I also modified a data pipeline in order to be able to convert raw URL data from Annotated-Document format to TensorFlow Example format using DocJoin and Flume

I researched and worked on a new method for data storage, optimal for a mass amount of queries,as an alternative to HBase, which only allows for infrequent lookups.

I also learned AWS Practitioning fundamentals and combined AWS services Athena, S3, and along with Hive to minimize storage costs.

Furthermore, I experimented with different data compression and storage formats in order to achieve minimum cost and at the same time boost query efficiency. This includes data compression, repartitioning, and index storage.