My research lies at the intersection of Hamiltonian lattice gauge theory, quantum simulation, quantum algorithms, and quantum information theory. I am broadly interested in developing formulations of many-body quantum systems that are both physically faithful and amenable to quantum computation, and in understanding the mathematical and computational structures that govern their complexity.
A central theme of my research is the Hamiltonian formulation of lattice gauge theories, especially in regimes where quantum simulation may offer access beyond standard classical methods. My recent work has focused on topological phases in (2+1)-dimensional lattice QED with Wilson fermions, with the goal of understanding how topological structure is realized in fully gauge-invariant Hamiltonian settings.
In particular, I have studied how the choice of fermion discretization affects the emergence of topological response. This work shows that the commonly used staggered-fermion Hamiltonian formulation is constrained by an exact time-reversal symmetry and therefore does not realize nontrivial topological phases, whereas Wilson fermions support topological regimes with nonzero Chern number. I have also worked on extensions of this framework to one- and two-flavor theories coupled to dynamical U(1) gauge fields, where the resulting phase structure includes topological response and gauge-invariant diagnostics such as many-body Chern numbers and current correlators.
More broadly, I am interested in Hamiltonian lattice formulations that preserve gauge constraints microscopically while remaining suitable for exact diagonalization, variational methods, and eventual quantum simulation.
I am interested in quantum simulation as a tool for studying many-body systems that are difficult to access with classical methods. Much of my recent work has centered on state preparation and variational approaches for strongly correlated systems, especially lattice gauge theories and coupled fermion-boson models.
One direction of this research studies VQE-based approaches to lattice gauge theory, where the aim is to construct circuit ansätze and observables tailored to Hamiltonian gauge theories with dynamical matter. This connects naturally with my work on Hamiltonian formulations of topological phases in lattice QED, where variational methods provide a route toward probing phase structure and response in gauge-invariant Hilbert spaces.
A second direction concerns quantum state preparation for many-body systems more broadly, including adiabatic state preparation for unbounded Hamiltonian systems and models such as the Hubbard-Holstein system. Here I am interested both in rigorous error analysis and in practical protocols for preparing low-energy states of interacting models.
On the algorithmic side, I am interested in the design of quantum methods for classically hard optimization and simulation problems. A current focus is continuous-variable QAOA, where the goal is to develop low-depth quantum optimization protocols for continuous problems that retain the flexibility and practical appeal of QAOA while extending it beyond discrete settings.
In this line of work, I study both algorithm design and complexity-theoretic questions. This includes identifying problem classes where continuous-variable quantum optimization may offer meaningful advantages, understanding the resources required for implementation, and clarifying the relationship between physically motivated continuous dynamics and more abstract query-complexity separations.
I am also interested in quantum information-theoretic questions arising in many-body physics and machine learning. In recent work on neural quantum states, I studied how information-theoretic quantities can be used to characterize the representational complexity of many-body wavefunctions and the expressive power of machine-learning ansätze.
In particular, this work develops scaling laws relating the growth of dephased or amplitude-based mutual information to the capacity required of autoregressive neural-network quantum states. These results help explain when neural-network representations can outperform more traditional tensor-network descriptions, especially for state families whose relevant correlations scale beyond area-law behavior.