Awesome Open MLOps

Overview

Awesome Open MLOps focuses specifically on open-source tools for Machine Learning Operations, providing alternatives to commercial MLOps platforms.

Open-Source MLOps Platforms

• MLflow: Complete ML lifecycle management
• Kubeflow: ML toolkit for Kubernetes
• ZenML: Extensible MLOps framework
• Metaflow: Netflix's ML infrastructure framework
• Kedro: Production-ready data science

Model Serving

• Seldon Core: ML deployment on Kubernetes
• KServe: Serverless ML inference
• BentoML: Build production ML services
• TorchServe: PyTorch model serving
• TensorFlow Serving: TF model deployment

Experiment Tracking

• MLflow Tracking: Open-source experiment tracking
• DVC: Data and model versioning
• Aim: Experiment tracking and comparison
• Guild AI: Experiment management

Feature Stores

• Feast: Feature store for ML
• Hopsworks: Data-intensive ML platform

Monitoring and Observability

• Evidently AI: ML model monitoring
• Prometheus: Metrics and monitoring
• Grafana: Dashboards and visualization

Workflow Orchestration

• Apache Airflow: Workflow automation
• Prefect: Modern workflow orchestration
• Argo Workflows: Kubernetes-native workflows
• Flyte: Cloud-native workflows

Data Versioning

• DVC: Data Version Control
• Pachyderm: Data lineage and versioning
• LakeFS: Git for data lakes

Pricing

All tools listed are free and open-source.