Vishwak Srinivasan

Vishwak Srinivasan is a graduate student in electrical engineering and computer science working at the intersection of sampling and optimization theory. He develops sampling methods inspired by optimization techniques that leverage geometric structure of constrained spaces, including preconditioned variants that replace expensive nonlinear subroutines with efficient matrix operations. A complementary line of inquiry explores how Hamiltonian dynamics—a staple in state-of-the-art sampling algorithms—can accelerate convex optimization, with Vishwak demonstrating that trajectory aggregation yields provably accelerated algorithms. He uses computational tools for rapid prototyping of complex algorithms, verification of analytical results, and visualization of geometric concepts that guide theoretical development. As a MathWorks Fellow, Vishwak will develop a computational toolbox implementing these algorithms with visualization capabilities for practitioners and researchers. Vishwak’s work has the potential to make advances in sampling and optimization theory more accessible to researchers across disciplines.