Services

Internships

  • Systems Research Intern, Microsoft Research Redmond, September 2017-December 2017, Mentors: Sameh Elnikety, Ricardo Bianchini
    • Azure Benchmark Suite: We collected highly running workloads of Azure Compute and packed them into an easy-to-use and representative benchmark suite. We used PowerShell scripts and C# to implement AzureBench that can be used for many purposes such as performance analysis as well as power and resource provisioning. Using this benchmark suite, Azure will be able to assess the impact of any platform changes on the performance of its main workloads.
  • Systems and Infrastructure Intern, LinkedIn, June 2017-September 2017, Mentors: Jean-Francois Im, Ravi Aringunram
    • Design and implementation of a load-aware segment assignment strategy for Pinot: As of my second and main project I was working on how segments (please see Pinot documentation) should be distributed among Pinot servers to get lower latency and balanced server utilization. I implemented a generic load aware segment assignment strategy, and we have extended this work by proposing EASY in the ICDCS'18 conference.
    • Implementation of HAVING clause for Pinot SQL language: Pinot is an open source OLAP data store that is being developed and used at Linkedin. During this project, I wrote about 5K lines of Java codes (including unit and integration tests) to implement HAVING clause (https://github.com/sajavadi/pinot).