This year was the first time I felt I completed the ho so well defined(?) data science "life cycle" - what a ride!
In this talk I'll describe the interesting question we tried to solve predicting whether a task we present to a crowed through our crowedsource app will be successfully completed on time. Given the fact crowed is an ever changing workforce and our tasks are flexible I'll try to use this challenge to take you through the bottlenecks, valleys and picks involved in the cycle, from a data scientist perspective. Hopefully, it won't be just a support group for frustrated DS but also help other individuals (product, exec, dev, PMs to name a few) to understand better the process of working with and aside to data science teams.