HPC Engineer - Randstad USA
Beltsville, MD 20705
About the Job
Randstad Federal is seeking an HPC Engineer for a program supporting NASA
location: BELTSVILLE, Maryland
job type: Contract
salary: $70 - 80 per hour
work hours: 8am to 4pm
education: Bachelors
responsibilities:
An individual at this skill level should have demonstrated extensive experience working with common HPC batch schedulers e.g. (PBS, Slurm, or Moab/Torque) while contributing to the support of users of HPC resources on the various issues they might have getting applications to run efficiently. This individual should demonstrate experience installing, maintaining, and upgrading HPC systems. The individual, along with the entire HPC team, will be engaged in the day-to-day operations and support of the HPC resources. Activities may include system patching, OS upgrades, deploying new systems, writing scripts, and troubleshooting system issues on the HPC system. The ability to interact with users to determine symptoms, and then reproduce their issues to isolate the causes is critical skills for this work. There will also be activities in testing, benchmarking, user tool scripting, and analyzing trouble tickets to find patterns indicating system or user education issues.
Duties and Responsibilities:
- Designs, deploys and maintains HPC clusters with over 2000+ nodes with InfiniBand, 100+ petabytes of data storage in production.
- Write and shepherd scalable feature designs through the entire software development process, from requirements and use cases to release
- Designs and develops scripts for system administration, monitoring and usage reporting.
- Modify existing software to correct errors and/or improve performance
- Designs and develops scripts for system regression test and performance (file systems (Luster), scheduler (PBS), interconnect (HDR/NDR, Slingshot, ), high
- availability, etc.).
- Troubleshoots, isolates and resolves application, system and other technical problems (hardware, software, and network).
- Understands research use cases, researches and deploys new technologies, defining cost, performance and other trade-offs.
- Manages and maintains tools for configuration management (HPCM, Ansible & GIT), resource management, scheduling and all necessary aspects of HPC in accordance with best practices.
- Researches, deploys and manages networking and security infrastructure, including development of policies and procedures.
- Assists in developing and writing proposals and publications.
- Creates and provides clear documentation.
- Mentoring junior staff and cross training peers
- After hours/weekend support as required
Moderate Supercomputing System Administration that contributes to:
- Day-to-day operations of the Linux HPC clusters and storage systems
- Proactive monitoring, analyze, and correct system issues
- Development of scripts to automate repetitive tasks or tools to enhance support of the HPC systems
- System performance analysis and tuning
- Building, installing, and supporting user-requested software
- Supporting evaluation and assessment of new HPC technology
- Resolving user report issues and manage support tickets requests in Remedy
Requirements:
- Bachelor's degree in computer science or related field
- strong computer science background with in-depth systems-level knowledge in operating systems and networking
- Solid understanding of the software development process, including requirements, use cases, design, coding, documentation and testing of scalable, distributed applications in a Linux environment
- A minimum of 5 years' experience working with HPC systems and scheduling software (PBS, Slurm, or Moab/Torque)
- A minimum of 10 years of experience developing system software in heterogeneous, multi-platform HPC environments
- strong ability to analyze, debug and maintain the integrity of an existing code base
- Demonstrated equivalence of 5 years of Linux/UNIX user support experience and hands-on experience with administration of Linux systems