Systems for Managing the Machine Learning Development Workflow
Given the effectiveness of machine learning at solving many real-world problems, a major portion of modern software systems is made up of ML models as key system components. However, the plethora of available ML development tools made the process of developing ML applications increasingly complex in recent years. Developers are faced with a seemingly infinite number of possible choices and each one has the potential to result in poor performance. The ease.ml system is designed to help ML developers manage their workflow by focusing on key pain points and leveraging principled methods for guiding their decision making process.
Here is the content of my research.