SQLBits 2019

Deploy ML models faster with Data Science DevOps

In this session I will show you how to apply DevOps practices to speed up your development cycle and ensure that you have robust deployable models. We will focus on the Azure cloud platform in particular, however this is applicable to other platforms
As Data Scientists we are great at machine learning, statistical modelling, visualising data and using data to tell a story. What are we not so good at? A lot of the core skills required in traditional software development. If you answer no to any of the following you need to attend this session.
  • Do you source control your models?
  • Do you test your models?
  • Is the percentage of models deployed in production less than 10 per cent?
  • Did you deploy the model?
In this session I will show you how to apply DevOps practices to speed up your development cycle and ensure that you have robust deployable models. We will focus on the Azure cloud platform in particular, however this is applicable to other cloud platforms.