Nov 7 – 9, 2022
Prague, Czech Republic
Europe/Prague timezone

Reduce Time-to-Market and Capital Spend using Software Operators

Nov 8, 2022, 11:30 AM
Karlin 3 (Hilton Prague)

Karlin 3

Hilton Prague

25 Minute Talk Data Science & Data Platforms


James Beedy
Matheus Tosta


This presentation addresses long lived challenges in High Performance Computing (operations overspend and decreasing time-to-market) and demonstrates how the Omnivector Slurm Distribution is a means to solve these problems. James and Matheus will outline the use case for software operators in HPC, describe the composition of the Omnivector Slurm Distribution, what problems are solved using OSD, and provide a demo of deploying, submitting and evaluating an HPC workload.

Session author bios

James Beedy (Phoenx, AZ USA) is the founder of Omnivector and is an advocate of open source software. Working with Juju and MAAS technologies since their inception, James has a long lived presence in the Ubuntu community and thorough knowledge of Ubuntu ecosystem technology and its application in solving real world problems.

Matheus Tosta (Florianopolis, Brazil) is the platform team engineering lead at Omnivector. Developing software since 2018, Matheus started his career in college pursuing an engineering degree, and now works with Omnivector to deliver HPC products to worldwide organizations.

As a software engineer, Matheus is passionate about building machine learning applications, DevOps initiatives, and contributing to open source projects.

Outside of his professional life, Matheus is a lover of his family and a proud father. He enjoys a myriad of FPS games and outdoor sports.

Level of Difficulty Intermediate
Social Media @omnivectorhpc, @matheushent

Presentation materials

There are no materials yet.