Nowadays, computing platforms use a mix of different hardware technologies, as a way to scale application performance, resource capacities and achieve cost effectiveness. However, this heterogeneity, along with the greater irregularity in the behavior of emerging workloads, render existing resource management approaches ineffective. In the first part of this talk, I will describe how we can use machine learning methods at the operating system-level, in order to make smarter resource management decisions and speed up application performance. In the second part of the talk, I will present how we can accelerate certain components of such systems using visualization and computer vision methods. Finally, I will conclude with my vision of coupling machine learning and computer vision at the system-level and present open questions that make this research area exciting to work on!