ML Framework Solution

Production level solution for automating training of already developed models

Architecture Diagram

Description

• Easy to setup Kubernetes solution which provides a ready to use environment for training complex ML algorithms.
• High Availiblity API service for fast training and prediction of ML models.
• Resource consumption is optimized by on demand resource allocation.
• Can be easily integrated to pre-existing automation solutions.
• Customized kafka cluster with secure communication protocols.
• Highly scalable and language agnostic.

Setup Instructions

• Each component of this solution has it's own README file containing the description, setup, and usage.
• Follow the Readme files and setup these components in the following order.
  • kafka-setup (KAFKA CLUSTER)
  • ftp-server
  • model-trainer
  • kube-deployer
  • prediction

Usage

• Then take the test-user files and upload it to the SFTP folder of one of the user.
• Execute the training and prediction through the frontend application or the API service provided.