Decentralized AI with Scaleout Studio

Scaleout Studio manages the full life cycle of multi-institution cross-silo federations. Focus on alliance governance and let Studio manage the software and infrastructure.

One-click deployment of FEDn networks on K8s
Onboard clients and handle end-to-end security
Serve models and create custom dashboards

Scaleout Studio is a cloud native solution that lets you as a data scientist prototype, scale-up and deliver decentralized AI solutions in production.  A unique UI and best-of-breed developer tools supports alliance life cycle management from initial model preparation to client onboarding to global model delivery. Studio removes the distributed system complexity inherent in federated learning and lets your team focus on model development and governance. 

When should you use Studio: 

  • If you want to rapidly experiment with federated learning without the complexity of learning how to deploy and manage a distributed network.
  • If you want to build production grade federated learning systems across multiple institutions or cloud regions.
  • If you want to collaborate with others on development and governance.
  • If you want to scale up you federated learning pilots to production.
Overview of the technology stack. Our federated learning framework FEDn is integrated with state-of-the art storage, developer, and ModelOps tools to create an easy-to-use environment for the lifecycle of FL applications.

Federated learning

Studio uses the FEDn federated learning framework under the hood. FEDn is an open-source, modular and machine learning agnostic framework developed by Scaleout Systems in collaboration with research groups at Uppsala University, Zenseact, AI Sweden and others.

FEDn enables highly scalable cross-silo and cross-device FL use-cases over FEDn networks.

FEDn lets you seamlessly go from local development of a federated model in a pseudo-distributed sandbox to live production deployments in distributed, heterogeneous environments. Three key features of FEDn:

  • Use any ML-framework and any language for clients
  • Highly scalable through a hierarchical federated learning approach
  • Built for production scenarios
Resources

Get in touch to learn more!

Enter your email below

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.