Skip to content

What Is a Workspace?

A Workspace is a comprehensive virtual environment designed for managing the entire lifecycle of machine learning operations (MLOps) projects. In a single, secure setting, it offers tools for isolation, resource allocation, model deployment, and more.

Core Capabilities

1. Isolation and Security

  • Secure Boundaries – Each project operates in its own environment, preventing conflicts and safeguarding sensitive data.
  • Access Control – Manage user permissions within a workspace using Workspace Access. For broader governance, see User Governance.

2. Resource Allocation

  • Computational Resources – Distribute CPU, GPU, and memory to different projects within the workspace.
  • Administrative Control – Tenant admins can manage and monitor resource distribution.
  • Learn More – See Resource Allocation Documentation for details.

3. Usage Monitoring

  • Resource Utilization Tracking – Monitor CPU, GPU, and memory usage across experiments.
  • Optimization Insights – Identify bottlenecks and optimize overall resource allocation.

4. Model Management and Deployment

  • End-to-End Lifecycle – Oversee the development, validation, and production deployment of ML models.
  • Performance Monitoring – Keep tabs on model performance and quickly roll out updates.

5. Job Management

  • Efficient Workflows – Jobs are responsible for model training, and most importantly distribute training accross several nodes.
  • Timely Delivery – Automate tasks to ensure on-time completion of key deliverables.

6. Experiment Tracking

  • Version Control – Log and track results from various experiments.
  • Comparisons – Compare performance metrics across different models or code versions.

7. Dataset Management

  • Data Integrity – Store, access, and version datasets consistently.
  • Collaboration – Share and maintain datasets across projects without duplication.

Next Steps