About 80% of AI projects never make it into production.
Why do they fail? Top factors are misaligned objectives, lack of collaboration between business stakeholders and data scientists, and an inability to leverage the collective knowledge and talents of the entire “team”, including Data Scientists, Data Engineers, ML Ops, etc.
These problems are not unique. Agile practices have become the de-facto approach to deliver software applications effectively. Can we adapt them for data science projects?
It is not that easy. Most data science teams experience problems adapting Agile due to Data Science specifics. For example, they struggle to produce a valuable increment at the end of the sprint, sometimes newly discovered information can ruin a sprint plan right in the middle of it, etc.
We need to tailor Agile Practices to allow for Data Science specifics.
In this talk, we will explore collaborative techniques that guide data science teams in their agile adaption. We will discuss how to come up with nice and clear product hypothesis, how to prioritize them using ICE/RICE method, how to decompose huge AI Epics into a small and easy to validate data science hypothesis, and how to effectively manage work using Kanban and Scrum approaches.