What is PaletteAI?
PaletteAI is a self-service platform for data, AI and ML teams to deploy and operate AI and ML application stacks on Kubernetes quickly, repeatably and consistently, while delegating the underlying infrastructure to platform engineering teams.
Deploying hardware-dependent applications for AI and ML requires extensive knowledge of Kubernetes, GPUs, networking, storage and other infrastructure that data science teams that want to deploy these frameworks aren't experts in. With PaletteAI, platforming engineering teams take care of the infrastructural aspects so that the data science teams only have to deal with what actually matters to them.
Catalog of AI and ML Application Stacks
PaletteAI comes with a catalog of turnkey AI and ML application stacks to get started quickly. These profiles can be customized where needed to suit specific needs, and teams can create their own profiles to build custom, in-house stacks. With these application deployment profiles, teams can quickly and repeatably deploy many instances of applications to fit their team's data workflows and needs, increasing velocity and productivity and reducing time spent on configuring and deploying AI and ML stacks.
What is Under the Hood?
PaletteAI leverages Mural for application configuration and deployment, so that deployed instances of an application are consistent across clusters and environments. Mural uses WorkloadProfiles to capture its configuration and deployment specifics as YAML-based Kubernetes custom resources. Each workload is a deployed instantiation of a particular workload profile and deployed applications are tied to a versioned WorkloadProfile, enabling Mural to do lifecycle management, reconciling revisions of the profile with the deployed application.
Additionally, PaletteAI can leverage Spectro Cloud's Palette to create the right infrastructure, so that data science teams have the right Kubernetes cluster ready to provision their AI and ML stacks. PaletteAI can also use existing clusters.
Who Can Benefit From PaletteAI?
PaletteAI provides benefits to data science teams who want to deploy AI and ML application stacks, and platform engineers who maintain Kubernetes environments.
Data Science Teams
With PaletteAI, Data Science Teams enjoy the repeatability and consistency of codified application deployments, so that data scientists can focus on what matters to them instead of being bogged down with complex infrastructure configuration and application deployment complexities. Whether initial deployment or ongoing lifecycle management, PaletteAI makes sure applications are kept up-to-date.
Platform Engineering
Platform Engineering teams use Palette's Cluster Profiles to tailor provisioned Kubernetes clusters to AI and ML applications with the right hardware (GPUs, storage, networking) and cluster configuration to ensure optimal performance, while still meeting security standards and technical platform requirements. This declarative approach makes life easier for platform engineering teams with consistency, repeatability, and all the enterprise-grade controls and governance they need - for both training and inferencing clusters.
Next Steps
Learn more about PaletteAI and how it can improve your AI and ML application stack deployment experience. Review the architecture to understand how PaletteAI leverages Palette and Mural to deliver a seamless experience for end-users.