Prof. Felix Biessmann | Code Reviews, Testing and Documentation - Improving Processes in Data-Driven Software Development

Software systems in industrial applications as well as in research are becoming increasingly data driven: Many components leverage data to increase the degree of automation or gain insights from data. While there are standard processes for improving the quality of classical software, these tools often cannot be directly applied to software components that learn from data. In this talk we will revisit some best practices for classical software development and we will discuss some more recent approaches to improve the quality of modern data-driven software systems that can help to reduce technical debt and increase replicability of research as well as responsible usage of data-driven software.