In this talk, I’ll discuss my experiences in practice, the ongoing frustrations, and the challenges being faced in modern machine learning and cloud systems, with a focus on how software engineering can help us meet those challenges. In particular, I’ll describe the challenges of building real-world applications that involve machine learning and how practice is evolving to meet these challenges. This includes assurance, architecture, and practice. I’ll also explain how the widespread use of the cloud constrains design and the benefits and weaknesses of these constraints. I argue that software engineering has a role to play, perhaps more than ever, as modern practice evolves.
Watch the video.