Prasanna Balachandran, assistant professor of materials science and mechanical & aerospace engineering, is one of 20 rising stars identified by the editors of Computational Materials Science. The initiative recognizes the accomplishments and promise of researchers who are still in the early stages of their independent careers, within 10 years of receiving their Ph.D.s. The journal editors solicited nominations in December 2017 and invited 20 finalists to contribute a short review article to a virtual special issue of Computational Materials Science published August 2019. Prasanna's review article, “Machine Learning-Guided Design of Functional Materials with Targeted Properties,†presents results from two case studies that identified promising candidates for new materials with desired qualities such as better magnetic memories, high-temperature electric polarization, and the ability to change form without changing temperature. These material properties are key to low-power electronics and efficient energy conversion. In one study, Balachandran's data-driven, computational framework combined experiment data, group theory and machine learning methods to rapidly screen for new compounds with these qualities. This is akin to “finding a needle in the haystack†problem. Aided by a computational modeling method called density functional theory, machine learning homed in on 242 out of approximately 3,100 compositions. This case study demonstrates how classification learning methods can rapidly identify targeted design space and efficiently guide computational materials calculations. A second case study sequentially guided experiments to more quickly discover materials that could form in calcium titanium oxide minerals, popularly referred to as “perovskites,†with desired electro-optical properties for use in electronic components and microtransducers. To achieve this result, Balachandran formulated a novel, two-step machine learning method. When sequenced with regression analysis, experiments and feedback for adaptive learning, the method led to the discovery of six new compositions with the desired structure and with promising properties. These case studies demonstrate the efficacy of classification-based machine learning methods in the search for new materials. Machine learning weeds out uninteresting compounds, saving time and effort. Adaptive machine learning approaches that include a closed feedback loop with experiments or computations can effectively navigate the vast search space and provide much needed validation for data-driven predictions.