Our story of agile data science delivery
Machine learning can provide a fast track to solve hard problems and a proof-of-concept can often be created quickly. However, the way to production can still be long, since machine learning creates new challenges not encountered in conventional software engineering.
We are researching methods to automate critical text classification tasks in GfK’s data production since three years. To operationalize these models, our team builds and runs a set of interdependent services to be consumed by a legacy production system.
Frank Rosenthal and Laura Hoyden will talk about their journey and the lessons learned, focusing on how to balance data science, technology and agile delivery to achieve a sustainable business impact.
Vorkenntnisse
Basic understanding of machine learning and software development
Lernziele
• Understanding the challenges and being able to avoid potential pitfalls when building and rolling out machine learning solutions,
• Being aware of the issues and potential approaches to automate and stabilize the long-term operation of the machine learning solution