My research integrates computational harmonic analysis, numerical linear algebra, and applied optimization to design fast, reliable methods for modern signal and data problems. My research develops scalable matrix and tensor methods—including low-rank modeling, randomized algorithms, and CUR/Nyström-type decompositions—together with rigorous guarantees for accuracy, stability, and robustness. I am particularly interested in structure-informed sampling and recovery, where the sampling strategy is designed to exploit underlying structure (e.g., tensor, graph, or hypergraph structure) to enable reconstruction from limited measurements. Another major theme is dynamical sampling from partial space–time observations of evolving processes, emphasizing identifiability, stable recovery, and robustness to noise, missing data, and outliers.