Skip to main content
Version: docs v25.02

Try Mellerikat Tutorial

Mellerikat is an MLOps-based platform that efficiently supports the entire process of AI solution development, deployment, and validation. This guide walks you through the steps to develop and verify an AI solution from start to finish using Mellerikat, along with the required environments. Below is a step-by-step overview of the process, with brief explanations for each stage.

Preparation Stage​

To use the Mellerikat platform, you first need a user account, which grants access to tools like AI Conductor, Edge Conductor, and Mellerihub.

Account Creation and Assignment​

  • Request Access: Submit your request via the Contact Page, detailing your purpose and needs, to receive credentials for the Try Mellerikat platform.
    • A Github token for downloading AI Contents (Note: Requires a separate Github account).
    • AWS Key for AI solution and Edge App registration.
    • IDs for Try platform’s AI Conductor and Edge Conductor.
  • Contact the Team: Request account creation through the Contact Page with specific details about your use case.
  • Verify Access: Use the provided credentials to confirm login to:

AI Solution Development​

ALO Installation​

Install ALO (AI Learning Organizer) in your local environment to prepare for AI model development. Refer to the installation guide for setup details.

Solution Development​

Prepare data, design, and test an AI model to create your solution. This step involves defining the problem, gathering data, and coding the model.

Solution Registration​

Upload the completed model to AI Conductor to register it as a solution, making it ready for operational use.


AI Solution Operation​

Solution Training Request and Validation​

Request training for the registered solution via Edge Conductor to ensure it’s operational-ready. Verify the training results to confirm functionality.

Edge App Installation​

Install Edge App on an edge device or PC to deploy the trained model.

Solution Inference and Validation​

Run real-time inference with the deployed model and check its performance via Edge Conductor.


Environment by Stage​

Here’s a summary of the environments required for development and validation:

StageEnvironment
ALO InstallationUser’s local analysis environment (PC, server)
Solution DevelopmentUser’s local analysis environment (PC, server)
Solution RegistrationTry-Mellerikat environment (AI Conductor)
Solution Training RequestTry-Mellerikat environment (Edge Conductor)
Edge App InstallationUser’s local analysis environment (PC, server)
Solution InferenceUser’s local + Try-Mellerikat environment (Edge Conductor)

This process combines the developer’s local environment with Mellerikat’s cloud infrastructure. Edge App can be installed on a personal PC or cloud setup as needed. After understanding the overall flow, refer to the detailed guides for each step linked throughout this tutorial.