Desafios na colaboração entre engenheiros de software e cientistas de dados em projetos de machine learning.

Abstract

Recent advances in Machine Learning (ML) have sparked considerable interest in integrating AI capabilities into software and services, making collaboration between data scientists and software engineers crucial yet challenging. This study investigates the challenges encountered in this collaboration in ML projects. Through interviews with professionals in the field, we identify critical issues such as knowledge gaps between disciplines, adaptations of Software Engineering (SE) practices for ML, and difficulties in model evaluation. We highlight the importance of early involvement of data scientists in defining software requirements, contributing to the successful development of ML systems. This study provides valuable insights for teams facing similar challenges in ML implementation.