Skip to main content

Machine Learning & AI

Machine Learning & AI tutorials help you build practical AI workflows on KakaoCloud, from model development, training, and serving to LLM applications and AI search.

The tutorials cover Kubeflow-based MLOps practices as well as vector search, RAG, and natural-language log analysis scenarios that combine OpenSearch with LLMs.

Tutorial structure​

This category is organized into the following series based on learning goals and use cases.

Kubeflow-based machine learning workflows​

Practice the full machine learning pipeline with Kubeflow, including model development, experiments, training, tuning, and deployment.

  • Kubeflow basic workflows Learn core Kubeflow components such as Jupyter Notebook, Pipeline, Katib, TensorBoard, and KServe, and understand the basic model development flow.

  • Kubeflow-based LLM workflows Configure LLM inference endpoints, fine-tune models, and build RAG applications in a Kubeflow environment.

  • Kubeflow-based traffic prediction model Build an end-to-end predictive modeling workflow using load balancer log data, including feature engineering, model tuning, serving, and pipeline automation.

OpenSearch-based AI use cases​

Use Advanced Managed Search (OpenSearch) with existing data and LLMs to implement AI-powered search and analysis features.

  • Vector search Integrate the OpenAI Embedding API with OpenSearch to build a semantic search environment.

  • RAG-based Q&A Retrieve documents stored in OpenSearch and pass them to an LLM to generate context-aware answers.

  • MCP-based log analysis Connect Claude Desktop with OpenSearch MCP Server to analyze operational logs in natural language.