Deploy and scale AI/ML models in production

Function as a Service (FaaS) provides maximum flexibility and extensibility, it is easy and straightforward to deploy AI pipelines in functions:
[features engineering -> model -> prediction transformations]

Support PMML, Scikit-learn, XGBoost, LightGBM, Spark-ML; ONNX, TensorFlow, Keras, PyTorch, MXNet, and even custom models.

Manage analytic assets: models, versions, scripts, Jupyter Notebooks, and datasets in one system. Monitor performance and healthy of deployments, and web service metrics, all shown in the visualization dashboard.

REST API of real-time scoring, batch scoring, and model evaluations in both Dev and Product modes. Support connections of a wide variety of data sources: CSV, HDFS, and various relational databases.