Databricks just announced its partnership with Microsoft on the open source MLflow project. Microsoft is going to be an active contributor and adding native support for MLflow in Microsoft Azure Machine Learning service. Databricks itself is the leader in Unified Analytics and was founded by the original creators of Apache Spark™.
MLflow is an open source platform for the machine learning lifecycle. From the moment Databricks unveiled MLflow in June 2018 at the Spark + AI Summit, the community engagement and contributions received have resulted into support needed for multiple programming languages and integrations with popular machine learning libraries and frameworks.
It’s not up to a year since the project was started and MLflow already has more than 500K monthly downloads, 40 contributing organizations and over 80 code contributors. This reaffirms the necessity for an open source approach in a bid to standardize the machine learning lifecycle across teams, tools, and processes.
Azure Machine Learning is a famous machine learning service which enables Azure customers to train, build, and employ machine learning models. In a bid to give customers maximum flexibility, Microsoft is in full support of open source MLflow in Azure Machine Learning. This implies that developers can make use of the standard MLflow tracking API to track runs and deploy models directly into Azure Machine Learning service.
“We’ve been thrilled by the contributions from Microsoft and the wider MLflow community. These contributions include support for multiple new programming languages, popular machine learning frameworks and services. It is inspiring to see the rapid adoption of the project. In terms of contributor count, MLflow achieved in 9 months what Apache Spark took 3 years to achieve. We have an aggressive roadmap for MLflow in 2019 and are excited to work with the community to expand the project,” said Matei Zaharia, chief technologist and co-founder at Databricks, and the real creator of MLflow and Apache Spark.
Rohan Kumar of Azure Data at Microsoft said, “We’re committed to the MLflow open source project, and leveraging and contributing to the innovations that are created in the community. We will continue to contribute innovations to make the machine learning lifecycle more efficient for Microsoft Azure customers.”