Overview
MLRun is an open-source AI orchestration framework (now part of McKinsey QuantumBlack) that automates data prep, tuning, validation and scalable serving on Kubernetes.
Key Capabilities
- Python SDK & serverless function runtime
- Hybrid/on-prem, multi-cloud deployment with auto-scaling
- Built-in feature store & drift monitoring
- Real-time model serving with GPU / TensorRT support
- Central dashboard for lineage, metrics and alerts
