Krishnakumar Balasubramanian

Vanderbilt University, University of California, Davis

Papers

Total Citations

H-Index

About

Krishnakumar Balasubramanian’s research bridges two seemingly distinct worlds: model-driven engineering for distributed real-time and embedded (DRE) systems, and high-dimensional stochastic optimization on Riemannian manifolds. His early seminal work on model-driven middleware introduced a paradigm that revolutionized how DRE applications—critical in avionics, telecommunications, and defense—are developed and provisioned. This foundational contribution, cited over 50 times, established a framework for automating the synthesis and configuration of middleware, enabling complex, resource-constrained systems to meet stringent timing and reliability requirements. More recently, Balasubramanian has advanced the frontiers of optimization theory. His 2022 paper on stochastic zeroth-order Riemannian derivative estimation tackles the challenging problem of optimizing functions over curved spaces (e.g., spheres, Stiefel manifolds) when only noisy function values are available—no gradients. By proposing novel gradient estimators, he opened doors for machine learning applications in geometry-aware settings, such as subspace tracking and matrix completion. With over 20 citations in just a few years, this work is rapidly gaining traction. Balasubramanian’s unique ability to contribute both to practical software engineering and to rigorous mathematical optimization underscores his versatility and impact.