Curated list of tools, frameworks, and resources for building, deploying, and monitoring machine learning systems in production.
- ML Pipelines & Orchestration
- Model Serving
- Feature Stores
- Experiment Tracking
- ML Platforms
- Monitoring & Observability
- Data Validation
- CI/CD for ML
- Infrastructure & Deployment
- Books
- Courses
- Kubeflow - ML workflow platform on Kubernetes
- Apache Airflow - Programmatic workflow orchestration
- Prefect - Modern workflow orchestration
- Dagster - Data orchestrator for ML pipelines
- Flyte - Workflow automation platform
- ZenML - MLOps framework for reproducible pipelines
- Metaflow - Framework for real-life data science
- MLflow Pipelines - MLflow's pipeline framework
- BentoML - Unified model serving framework
- Triton Inference Server - NVIDIA's production inference server
- TorchServe - PyTorch model serving
- TensorFlow Serving - TF model serving
- FastAPI - Python web framework for ML APIs
- Seldon Core - MLOps for Kubernetes
- Ray Serve - Scalable model serving
- Feast - Open-source feature store
- Tecton - Enterprise feature platform
- Hopsworks - Feature store + ML platform
- MLflow Tracking - Log parameters, metrics, artifacts
- Weights & Biases - Experiment tracking and visualization
- Neptune.ai - Metadata store for MLOps
- Comet - Experiment tracking and optimization
- DVC - Data and model version control
- Sacred - Experiment configuration and logging
- Kubeflow - ML toolkit for Kubernetes
- H2O AI Cloud - Enterprise AI platform
- DataRobot - Automated ML platform
- Valohai - MLOps platform
- DagsHub - ML platform with tracking and collaboration
- Evidently AI - ML monitoring and testing
- WhyLabs - AI observability platform
- Great Expectations - Data quality validation
- Prometheus - Metrics collection and alerting (for ML infra)
- Grafana - Monitoring dashboards
- NannyML - Post-deployment monitoring
- Alibi Detect - Drift detection
- Great Expectations - Data quality tool
- Pandera - DataFrame validation
- TensorFlow Data Validation - Data validation for ML
- Deequ - Data quality on Spark
- GitHub Actions - CI/CD with ML workflow templates
- GitLab CI - Built-in CI/CD
- Jenkins - Automation server
- CircleCI - CI/CD platform
- CML - Continuous Machine Learning
- Docker - Containerization
- Kubernetes - Container orchestration
- Helm - Kubernetes package manager
- Terraform - Infrastructure as code
- Pulumi - IaC with general-purpose languages
- AWS SageMaker - Managed ML service
- Vertex AI - GCP ML platform
- Azure ML - Azure ML service
- Designing Machine Learning Systems - Chip Huyen
- Building Machine Learning Pipelines - Hannes Hapke & Catherine Nelson
- Machine Learning Engineering - Andriy Burkov
- The ML Test Score - Google's ML testing framework
- Introducing MLOps - Mark Treveil
- Made With ML MLOps Course
- Full Stack Deep Learning
- DataTalks.Club MLOps Zoomcamp
- Chip Huyen's ML Systems Design
Contributions welcome! Submit a PR or open an issue.
MIT