Skip to content

Workspace User Interface

Workspaces Landing Page

Upon entering the platform, you’ll see a table of all existing workspaces. To create a new workspace, click New Workspace.

Workspace overview

Selecting any workspace opens its detailed view.

Inside a Workspace

By default, you’ll land on the Experiment Tracking page. A left sidebar provides quick access to various tools and features.

Workspace Sidebar

1. Tracking

  • Experiment Management – Log runs, compare models, and analyze results.
  • Detailed Logging – Capture metrics, parameters, and output data for each experiment.

2. Models

  • Model Registration – Easily register ML models for future reference.
  • Deployment Prep – Streamline the handoff from development to production.

3. Deployments

  • One-Click Launch – Deploy models to production with minimal hassle.
  • Live Monitoring – Track model performance and quickly roll out updates.

4. AB Testing

  • Evaluate and Compare – Run A/B tests on model deployments in OICM—split traffic, collect real‑time feedback, and identify the best‑performing model for GenAI and RAG workloads.
  • Lauch, Review, and Declare a Winner – configure, launch, and evaluate A/B test experiments in OICM—set traffic splits, choose feedback types, apply statistical thresholds, and declare a winner with confidence.

5. Jobs

  • Custom Training – Use your own training scripts and configuration files.
  • Resource Management – Monitor GPU/CPU usage for efficient execution.

6. LLM

  • Fine-Tuning & Benchmarking – Access pre-built tools to customize large language models.
  • Dataset Integration – Connect various data sources for thorough testing.

7. Annotations

  • Human Feedback – Incorporate user or team feedback into training data.
  • AI-Assisted Labeling – Speed up labeling tasks with AI-driven suggestions.

8. Blueprints

  • Reusable Templates – Save configurations, secrets, and workflows for future use.
  • Rapid Setup – Launch new projects quickly using existing best practices.

9. Datasets

  • Data Management – Create, version, and organize datasets.
  • Consistency & Integrity – Keep data consistent across all projects.

10. Notebooks

  • Built-In Environment – Leverage a Jupyter-based setup directly inside the platform.
  • Seamless Analysis – Explore data and run code without leaving OICM.

11. Usage Monitoring

  • Admin Oversight – Track resource utilization across users and workspaces.
  • Cost & Performance – Identify bottlenecks and optimize workload distribution.

12. User Governance

  • Fine-Grained Permissions – Manage roles, ownership, and access rights.
  • Workspace Control – Assign or revoke user privileges as needed.

13. Settings

  • Custom Configuration – Adjust workspace parameters to fit your workflow.
  • Global Policies – Manage security, resource limits, and environment settings.

Next Steps