What's In ModelOp?

ModelOp Center is AI/ML governance software that provides comprehensive oversight, control, and management for all AI/ML models across your enterprise, regardless of their origin, runtime platform, or consuming application.

ModelOp Center delivers complete model lifecycle management through core services including Model Manager, MLC Manager, and ModelOp Runtime, accessible via an intuitive UI with modules for inventory, orchestration, compliance, and operations.

Core Architecture

ModelOp Center is built on a microservices architecture deployed to Kubernetes, consisting of essential core services and optional integration services.

Core Microservices

Model Manager: Collects and persists all model metadata, technical assets, notifications, approvals, jobs, and test results. Backed by MongoDB, it provides the evergreen Model Inventory and detailed information for each model.

MLC (Model Life Cycle) Manager: Orchestrates and manages all processes throughout a model's entire lifecycle—from registration through promotion, production deployment, monitoring, and eventual retirement.

ModelOp Runtime: Language and framework-agnostic model execution environment supporting batch, streaming, and REST execution modes for models of all types.

Gateway: Provides secure entry point to all core ModelOp services.

Registry: Internal registration service enabling automatic discovery and monitoring of all microservices within ModelOp Center.

Reporting: Flexible service for generating reports from MLC Manager and other internal/external runtimes and services.

Document: Automatically generates documents using model metadata, test results, and templates.

Optional Microservices

Model Factory Plugins: Integration plugins for common model development environments (Jupyter, RStudio) enabling easy registration and management of models created within those platforms.

Model Execution (Runtime) Services: Integration services for various execution runtimes including Spark, SageMaker, and others.

Logstash: Output configuration for directing ModelOp service logs to your preferred destinations.

User Interface Modules

ModelOp Center's UI provides comprehensive access to all governance and operations capabilities through five major areas:

Executive Dashboard

The home screen delivers enterprise-wide visibility into all production models across environments, with KPI insights covering business, operational, data science, and risk metrics.

My Work

Personalized dashboard summarizing key risks, issues, and tasks requiring attention for models under your purview—your daily action list.

Inventory Module

Inventory Overview: View all models by deployment stage (Dev, QA, Prod), organized by Organization and Runtime, with model-specific notifications.

Inventory List: Complete Governance Inventory containing all Use Cases, Implementations (Models), and Monitors across the enterprise, regardless of model type, factory, or deployment state.

Learn more about Inventory & Management

Model Lifecycles Module

Model Lifecycles Overview: View active MLC processes, process instances, recent incidents, and issue trends over time.

Model Lifecycle List: Browse and manage all available model lifecycles within the system.

MLC Process Definitions: View process blueprints showing the workflow design, active instances, and any Camunda incidents.

MLC Process Instances: Monitor live automations running for specific models, tracking step-by-step execution and identifying technical issues.

Learn more about Model Lifecycle Management

Compliance Module

Compliance Overview: Executive dashboards displaying models by risk classification and business unit, lifecycle status, production compliance breaches, and validation expiration tracking.

Reporting: Generate detailed audit reports for deployed models at specific dates, downloadable as PDFs with complete deployment details.

Learn more about Monitoring & Reporting

Operations Module

Operations Dashboard: High-level view of running models, jobs, and open tickets, plus deployment status by runtime and recent notifications.

Jobs List: All jobs orchestrated by ModelOp Center, whether running on ModelOp Runtimes or external platforms (Spark, SageMaker).

Job Details: Complete job information including execution parameters, runtime location, timing, associated models, inputs/outputs, and logs.

Runtimes List: All registered runtimes (ModelOp and external) with current state, type, tags, stage, deployed models, and last activity.

Runtime Details: Detailed view of runtime status, configuration, engine details, deployed models, endpoints, platform libraries, and live performance statistics.

Deployments List: Running view of all deployments (Batch, REST, Streaming) across all environments and runtimes, filterable by state.

Notifications List: System-wide notifications for models, runtimes, jobs, and processes, filterable by type and severity.

Key Capabilities

ModelOp Center enables comprehensive model governance through three primary capability areas:

Inventory & Manage: Centralized inventory of all models with metadata management, use case tracking, and implementation details.

Orchestrate: Automated lifecycle workflows, governance processes, and deployment management across multiple environments.

Monitor & Report: Continuous monitoring, out-of-the-box tests, custom monitors, champion/challenger comparisons, executive dashboards, and automated documentation generation.

Getting Started

New users should begin with:

• ModelOp Center Terminology
• Getting Oriented with the ModelOp Center UI
• Getting Started with ModelOp Center

API Access

All features available in the UI are also accessible via the ModelOp Center REST API, enabling programmatic integration and automation.

‍