My research seeks to accelerate machine learning workflows by tackling the problem from systems and learning objectives. This involves developing efficient search methods that account for both objectives, evaluating trade-offs and accounting for constraints such as power budgets and task scheduling. The components that I am investigating include reduced precision computation and systems metrics in machine learning. The end goal of my work is to develop an end-to-end framework that provides recommendations to the end user, based on requirements, compute budgets, and problem constraints of the user.