Role

Intern, May 2019 - Aug. 2019

Website

Contributed to the development of an exabyte-scale distributed database. The internship was divided into two parts on two different teams to get a feel for different parts of the company.

On the first team, I worked on designing a testing framework to inject errors into system calls made by the database. This allowed the team to simulate failed hardware and improve fault tolerance. In addition, I worked on migrating a complex testing script from its archaic Bash form into Python, allowing easier upkeep and better integration with Ocient’s other Python libraries. Finally, I coupled output from testing suites with JIRA tracking to enable teams to quickly determine how tests perform on nightly builds.

The main project on the second team surrounded optimizing Ocient’s implementation of a distributed consensus protocol. Key features included optimizing for new nodes to catch up faster to an existing database, rejecting nodes that are slowing down the system, and allowing the database to be quickly rebuilt from an invalid state. The improvements not only provided faster database speeds for internal testing but also added powerful features of which clients can take advantage.