This research introduces a conversational AI system that transforms how humans interact with ML infrastructure through natural language.

Key Innovations:

• Hierarchical, modular architecture integrating specialized agents for ML workflow management
• KubeFlow Pipelines Agent for orchestrating ML pipelines through conversation
• MinIO Agent for streamlined data management
• Designed to support users with varying technical backgrounds

Engineering Impact: The Swarm Agent architecture provides a scalable framework for building conversational interfaces to complex MLOps systems, potentially reducing technical barriers and improving collaboration between technical and non-technical stakeholders in ML projects.

Towards Conversational AI for Human-Machine Collaborative MLOps