Seldon Core 2 is a modular framework with a data-centric approach, designed to help businesses manage the growing complexities of real-time deployment and monitoring for machine learning and AI models. Its data-centric and modular design ensures accurate, adaptable data, fostering confidence in models deployed in production at scale. The platform offers a flexible, platform- and integration-agnostic framework, enabling seamless on-premise or cloud deployments for any model or purpose, regardless of the existing tech stack requirements. Seldon Core 2 was developed to centralize data in machine learning deployments, enhancing observability for better understanding, trust, and iteration of current and future projects. It supports a wide range of runtimes, allowing teams to leverage pre-trained models from popular libraries like Triton (via ONNX), PyTorch, TensorFlow, TensorRT, MLFlow, Scikit, XGBoost, and Hugging Face, as well as custom developments. The platform facilitates seamless integrations with CI/CD pipelines, automation tools, and various ML tools, whether cloud-based, in-house, or third-party. Deployments are flexible and standardized, supporting environments such as GCP, Azure, AWS, RedHat OpenShift, or on-premise. Users can deploy traditional ML models, custom models, or Generative AI models—either as single models or complex applications—using a consistent workflow, and can mix models and components with both custom and out-of-the-box runtimes. The system is designed to increase productivity through improved workflows and more efficient resource utilization.

Features & Benefits

• Data-Centric & Modular Design
  • Ensures accurate, adaptable data and fosters confidence in models in production at scale.
• Platform & Integration Agnostic
  • Enables seamless on-premise or cloud deployments for any model or purpose regardless of tech stack requirements.
• Multi-Runtime Support
  • Allows teams to benefit from a broad range of pre-trained models, supporting Triton via ONNX, PyTorch, TensorFlow, TensorRT, MLFlow, Scikit, XGBoost, Hugging Face, and custom developments.
• Seamless Integrations
  • Connects with CI/CD, automation, and various ML tools (cloud, in-house, third-party).
• Flexible & Standardized Deployment
  • Deployable on GCP, Azure, AWS, RedHat OpenShift, or on-premise. Supports traditional ML, custom models, or GenAI as single models or complex applications with consistent workflows.
• LLM Module
  • Facilitates the deployment of popular GenAI models into production with capabilities designed to optimize and transform business operations.