This is a 6 month-1 year 1099 contract position working on a research project partnering with Lawrence Livermore National Laboratory to benchmark MapReduce performance when running on Lustre.
Primary Duties / Responsibilities
Will Modify a MapReduce implementation to be aware of data placement in a Lustre filesystem.
Devise data placement algorithms to optimize MapReduce performance when running on a Luster filesystem instead of HDFS.
Benchmark MapReduce performance in a variety of configurations on Lustre.
Write a research paper documenting the configurations and performance results.
Qualifications (Knowledge, Skills, Abilities)
Candidate must have expertise with
Requirements (Education, Certification, Training, and Experience)
Minimum Education: BS Degree in CS/CE prefer: MS or PhD in CS, CE or Parallel File Systems
Demonstrated experience in writing scientific research documentation
US Citizenship is Required
Candidate must be based in the US
Physical Demands / Work Environment
Ability and desire to work as part of a geographically-distributed team