This is an MLOps Radar. An ML project can be in two phases, either (A) exploration or (B) productionisation & iteration.
- Jupyter on local machine
- Jupyter on K8s multi-user
- Deepnote
- Google colab
- SageMaker Studio Lab
- GitHub Codespaces
- Databricks community
- Kubeflow Notebooks
- Vertex AI Workbench
- Einblick
- Weights & Biases
- MLflow
- Tensorboard
- Comet.ml
- Neptune.ai
- DVC
- SageMaker experiments
- Aim
Productionisation & Iteration phase
- SageMaker pipelines
- AWS Step Functions
- Metaflow
- Kubeflow pipelines
- Valohai
- Airflow
- TFX
- Bodywork
- ZenML
- SageMaker Feature Store
- Tecton
- Databricks Feature Store
- Feast
- SageMaker
- AzureML
- Vertex AI
- Vertex AI Vizier
- Anyscale
- SageMaker
- AzureML
- Vertex AI
- Kubernetes
- TensorFlow serving
- TorchServe
- SageMaker
- AzureML
- Seldon Core
- BentoML
- Anyscale
- Superwise.ai
- SageMaker model monitoring
- SageMaker clarify
- AzureML interpret
- AzureML studio
- Alibi Detect
- Alibi Explain