Towards Conversational AI for Human-Machine Collaborative MLOps

2025-04-16

AI Summary

High operational complexity of MLOps platforms (e.g., Kubeflow) impedes adoption by cross-disciplinary users lacking specialized expertise. Method: We propose a conversational LLM-based agent system featuring a hierarchical, scalable Swarm Agent architecture. It orchestrates multi-role agents with integrated RAG-enhanced knowledge retrieval, Kubeflow Pipelines (KFP) orchestration, MinIO-backed data/artifact management, and context-aware iterative reasoning—enabling the first end-to-end natural language interface for complex MLOps platforms. Contribution/Results: The system significantly improves task success rates and efficiency for non-expert users across ML pipeline authoring, data/artifact management, execution monitoring, and documentation querying. It reduces reliance on domain-specific technical skills and enables seamless participation of diverse stakeholders—including data scientists, engineers, and business analysts—in the full model development, deployment, and governance lifecycle.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: Conversational AI/Dialog Systems

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Search and Retrieval-Augmented AI: Agentic search

Abstract

This paper presents a Large Language Model (LLM) based conversational agent system designed to enhance human-machine collaboration in Machine Learning Operations (MLOps). We introduce the Swarm Agent, an extensible architecture that integrates specialized agents to create and manage ML workflows through natural language interactions. The system leverages a hierarchical, modular design incorporating a KubeFlow Pipelines (KFP) Agent for ML pipeline orchestration, a MinIO Agent for data management, and a Retrieval-Augmented Generation (RAG) Agent for domain-specific knowledge integration. Through iterative reasoning loops and context-aware processing, the system enables users with varying technical backgrounds to discover, execute, and monitor ML pipelines; manage datasets and artifacts; and access relevant documentation, all via intuitive conversational interfaces. Our approach addresses the accessibility gap in complex MLOps platforms like Kubeflow, making advanced ML tools broadly accessible while maintaining the flexibility to extend to other platforms. The paper describes the architecture, implementation details, and demonstrates how this conversational MLOps assistant reduces complexity and lowers barriers to entry for users across diverse technical skill levels.

Problem

• Enhancing human-machine collaboration in MLOps via conversational AI
• Simplifying ML workflow management through natural language interfaces
• Bridging accessibility gaps in complex MLOps platforms like Kubeflow

Innovation

• LLM-based conversational agent for MLOps
• Extensible Swarm Agent with modular design
• Hierarchical agents for pipeline and data management