Swiss Python Summit 2018 - 12-Factor Apps for Data Science with Python

Posted on February 16, 2018 in python talk

Heroku distilled their principles to build modern cloud applications. These principles have influenced many of our design decisions at Blue Yonder to build a data science platform.

Heroku distilled its principles for building modern cloud applications—maximizing developer productivity and application maintainability—into the 12factor.net manifesto. These principles have influenced many of our design decisions at Blue Yonder.

While our data scientists focus on machine learning models and statistics, we want to free them from concerns about technical operations such as network equipment maintenance, operating system updates, or hardware failures. To relieve our data scientists of these tasks, we have invested in a data science platform.

This talk provides insight into how we use Apache Mesos, Devpi, Graylog, and Prometheus/Grafana to provide a developer-friendly environment that lets data scientists build distributed Python applications without managing servers or scaling.