A major direction of my work is the integration of machine learning with first-principles plasma modeling. In funded work on multipactor, I pursue supervised ML frameworks for predicting susceptibility and threshold behavior, together with data-driven modeling of secondary electron emission under realistic surface conditions for accelerator RF components. In newer work, I am extending this agenda toward physics-aware AI for kinetic plasma discovery workflows, including magnetic reconnection analysis in fusion-relevant plasma. My interest is not in black-box prediction alone, but in interpretable ML that remains tied to physically meaningful observables and can accelerate scientific understanding as well as design. 

Methodologically, my work is built around multi-fidelity computational plasma science. Depending on the problem, I use analytical models, Monte Carlo and statistical methods, reduced fluid descriptions, and full particle-in-cell simulations, with each level used to extract different physical insight and computational leverage. I am particularly interested in connecting these approaches in a way that preserves physical interpretability while enabling scalable exploration of parameter space, realistic device configurations, and challenging regimes where purely analytical theory becomes intractable. 

A major part of my research focuses on gas discharge initiation, plasma breakdown, and transient plasma evolution under externally driven conditions. I develop reduced-order but physics-rich models for discharge formation and nonlinear plasma response, with emphasis on drift–diffusion–Poisson transport, dielectric charging, plasma–surface interaction, and external-circuit coupling. I led the development of the PASCHEN-1D framework, a unified platform for simulating breakdown and discharge transitions in DC, pulsed, and emission-driven systems, including nanosecond pulsed dielectric-barrier discharges, Townsend-to-glow transitions, Paschen behavior, and ultrafast photoemission-induced plasmas.

I also study multipactor and related RF-driven electron multiplication processes relevant to particle accelerators, high-power microwave devices, and vacuum RF systems. My recent work has examined how engineered waveforms and multi-frequency excitation modify susceptibility boundaries, mode structure, deposited power, and asymmetry of conductor loading. This line of research treats multipactor not only as a breakdown problem, but also as a controllable nonlinear instability whose behavior depends sensitively on geometry, excitation waveform, and surface-emission physics. 

My research is motivated by problems in particle accelerators, high-power RF systems, pulsed-power and vacuum-electronics environments, and increasingly in laboratory and fusion-relevant plasmas. Across these domains, my broader goal is to build computational tools that can both explain complex transient behavior and support the design of more reliable and controllable plasma-based technologies.