Azure Stream Analytics now supports high-performance, real-time scoring by taking advantage of custom pre-trained machine learning models that are managed by Azure Machine Learning service and hosted in Azure Kubernetes Service (AKS) or Azure Container Instances, using a workflow that doesn't require users to write any code. Build custom models by using any popular python libraries such as Scikit-learn, PyTorch, TensorFlow, and more to train your models anywhere including Azure Databricks, Azure Machine Learning Compute, and HD Insight. Once deployed in AKS or Container Instances clusters, you can use Stream Analytics to surface all end points within the job itself.
Sign up for a private preview of this feature now.