After an initial data-science proof-of-concept is completed, it often needs to be professionalized and deployed to production. Underestimating this step often leads to higher maintenance and slower time-to-market of new features. In this talk we will cover a set of practical software-engineering best-practices for industrializing a machine-learning model to help minimize such problems.

After this talk, you will know how to develop and run a professional data-science application in production like a pro. Among others, we will cover topics like:

• optimal deployment pipeline (automate all the things),
• Docker (no more "Works on my machine")
• team-collaboration (pull-requests for the win, no more bus-factor of one),
• application & log monitoring (a.k.a "those graphs on that monitor look impressive. Is that real-time?"),
• configuration (one ring to rule them all) ,
• REST APIs ("I don't care how it's implemented, just give me the endpoint"),
• and more (e.g. useful Slack notification)

The talk is based on my 10-year experience as a software-engineer and all the stuff I have learnt about developing and deploying software applications during this time.

This talk is suitable for everyone responsible for running a data-science model in production environments.