Spell MLOps Platform Launches ‘Spell for Private Machines’ to Streamline DevOps and Foster Deeper Team Collaboration for Enterprises

Print Friendly, PDF & Email

Spell – a leading end-to-end machine learning platform that empowers businesses to get started with machine learning projects and make better use of their data – announced its new Spell for Private Machines integration. With Spell for Private Machines, enterprise teams that are spearheading machine learning projects will be able to use their privately owned GPUs or CPUs alongside cloud resources for experimentation, results and collaboration, reducing time, money and resources usually spent in-house.

The Spell platform is built for team collaboration. With flexibility across computing options, this integration allows teams to scale from private machines to cloud computing when needed, allowing users to maximize the capacity of existing machines. Spell for Private Machines also takes care of the environment and dependency management of machine learning projects on your private machines and automatically tracks and visualizes your machine learning metrics.

“With our infrastructure for ML experimentation and deployment we can work on top of any kind of hardware,” said Serkan Piantino, co-founder and CEO of Spell. “Spell for Private Machines is a great option for teams that have invested in their own machines but could benefit from the flexibility of on-demand resources in the cloud. Ultimately, this feature will increase efficiency and productivity among teams and projects by automatically routing them to the fastest hardware in real time. And with Spell, teams will still have their Machine Learning work organized and available on one collaborative platform, no matter what hardware it was trained on.”

Spell for Private Machines allows team leads to track and monitor their team’s machine learning experiments seamlessly by having their private machines and cloud machine learning projects in one central location. Machines can be added and removed as needed and all traffic is encrypted and secure. All experiment results, checkpoints, and training outputs are automatically saved and accessible via the Spell web interface. After integration, scaling experiments from a private machine to an AWS v100x8 GPU cloud machine can be done in a few simple clicks. 

Sign up for the free insideBIGDATA newsletter.

Speak Your Mind

*