A confluence of advances has led to an inflection in our ability to collect, store, and harness large amounts of data to make predictions and guide decision making. After discussing recent developments and trends in machine learning, I will present several representative efforts on learning and inference, including projects that have transitioned from the research lab into the open world. I will first describe work to build and deploy predictive models that infer and forecast traffic flows in greater city regions. Then, I will present research on learning and fielding predictive models in healthcare. Finally, I will review efforts to glean insights from large stores of behavioral data, covering projects that leverage anonymized streams of data gleaned from cell towers, search engines, and social media.