How can architecture support and enable Data Science?

How can architecture support data science? Data scientists tend to use ad hoc approaches and programming languages and frameworks in which they are most skilled. In such a fast moving field, having complete freedom of choice with programming languages, tools, and frameworks can make for quick advances, but how do we get models from concept to production in a way that is repeatable, robust and delivers value? How do we avoid the wild west of desktop data science?

Join us from 6pm on Tuesday November 19th at 1 New York Street, Manchester. ECOM are kindly hosting us in their event space and are also providing pizza, beer, soft drinks and more.

As always this will be a group discussion. So please come and join in. This month we'll be discussing questions like:

How can architecture support data science?
How should data science fit into lean and agile teams?
What does good look like for model deployment?
What do Data Scientists really want from Architecture?