CloudML

Get Started

No-Code AI Analysis and Machine Learning Platform for Effective AI Application

CloudML is a machine learning platform that manages AI model lifecycle based on the MLOps pipeline. It provides a proven solution for AI application, reflecting the latest trends such as No-Code AI, integrated collaborative environments, easy lifecycle management, and Generative AI based Copilot.

Overview

  • No-Code AI

    CloudML provides an AI/ML analytics tool that can manage AI/ML lifecycle with low-code or no-code. By visualizing the entire analytics process, from data collection, storage, processing to analysis, modeling, deployment and monitoring, data analysis process becomes faster and more simple.

  • Integrated Platform for Collaboration

    By integrating SCP Container Registry by cluster type, a shared analysis environment is configured immediately, allowing for seamless collaboration. The shared analysis environment can also be updated/reconfigured based on needs.

  • AI Lifecycle Management

    CloudML provides a pipeline that can register Jupyter Notebook and workflow modeler as steps. It enables optimal analysis jobs for each lifecycle of the AI model. Log monitoring is provided in real-time during AI model learning, giving users an ata-glance view of the experiment indicators of each step and allowing them to track learning history.

  • Copilot Based on Generative AI

    Automatically generates code based on natural language queries. Maximize productivity and efficiency by recommending customized workflow functions. In addition, it provides a built-in Large Language Model(LLM) for easy copilot use without external LLM connection.

01

04

Service Architecture

CloudML
  • AI Lifecycle Management MLOps Pipeline
    • Model Development : Workflow Modeler/ Notebook
    • Training : Experiments / Verification
    • Deployment : Model Serving
  • Automation : Guided Analytics
  • Analytics Enviornment : Sandbox/ GPU Scheduler/ Custom Kernel
  • Visualization: Report/ Analytics App
  • Copilot
  • Kubernetes Engine

Key Features

  • Provide environment for model development and learning
    1. Jupyter Notebook-based model development environment
    2. Provide an option to select image by Notebook instance type (Including one built-in image)
    3. Custom Image can be utilized through SCP Container Registry linkage
  • GUI-based convenient analysis environment
    1. Provides GUI-based workflow with drag & drop
    2. Visualize input data for easy verification of data by field
    3. Utilize graph to visualize output data that is easy to understand
    4. Provides AutoML function
  • Pipeline management and tracking
    1. Check the status and execution history by step
    2. Monitor in real-time cumulative log files during step execution
    3. Visualize and monitor experiment metrics(accuracy, loss, etc.) real-time
  • Manage and compare MLflow based experiments
    1. Manage entire experiments and view detailed information by model
    2. Compare and analyze past experiment history of models for experiment management
    3. Model Verification function for model verification
    4. Provide Model Registry for managing and distributing model image

Let’s talk

Whether you’re looking for a specific business solution or just need some questions answered, we’re here to help

Get in Touch