Computational Physics & Scientific Machine Learning
I’m an undergraduate researcher in computational physics working on machine learning methods for physical systems. I focus on learning-based approaches to simulation and modeling in high-energy physics and quantum many-body systems. I also build interactive physics demos and create visualizations.
Research
Developed a Fourier Neural Operator to model nonequilibrium dynamics in strongly correlated systems. Trained on small lattices, the model demonstrates zero-shot generalization to systems exceeding 1024×1024, producing physically consistent predictions far beyond the tractable regime of conventional solvers.
Developed transformer-based models for reconstructing jet kinematics in simulated collider data. Investigated missing momentum signatures associated with dark matter production using a full simulation pipeline (MadGraph, Pythia8, Delphes) on GPU-enabled HPC systems.
Projects
A real-time 1D solver for the time-dependent Schrödinger equation using a fourth-order Runge–Kutta integrator, allowing users to specify arbitrary potentials and simulate the resulting wavefunction dynamics.
An interactive 2D sandbox for simulating vehicle motion under realistic force models, including acceleration, tire friction, and lateral slip, with emergent behaviors such as skidding.
A classical particle system sandbox, modeling simplified baryon interactions using the Reid potential and attractive and repulsive electric forces, assuming aligned spins.
An interactive training tool for rapid chessboard coordinate recognition featuring timed drills, immediate audiovisual feedback, and performance tracking through streaks and response-time statistics.
A minimal, distraction-free text editor with a clean monospace interface, offering real-time document statistics, import/export functionality, theming, and custom markdown-like syntax highlighting.
A lightweight password generator that produces high-entropy, human-typeable passwords with structured randomness, balancing usability and security.
GLSL Fragment Shaders
A real-time interactive Monte Carlo spectral path tracer rendering thin-film interference on a reflective glass sphere, featuring rapid sample convergence, ACEScg color management, and a custom lift/gamma/gain-inspired grading pipeline.
A real-time volumetric path tracer visualizing atomic orbital probability densities as participating media, with absorption and scattering behavior modulated by the wavefunction’s phase.
An interactive 2D visualization of a hydrogen 1s orbital under a uniform external electric field, computed via time-independent nondegenerate perturbation theory and rendered with density gradients and contour lines.
A real-time landscape renderer procedurally generating island terrains, featuring reflective and refractive water, dynamic wave motion, and atmospheric scattering effects.
A real-time visualization presenting a grid of 2D slices of the hydrogen wavefunctions, designed for intuitive exploration and educational presentation of orbital structure.
A real-time Mandelbrot set visualization with interactive lighting control, producing a shaded, quasi-3D appearance.
Contact
I’m interested in graduate study at the intersection of physics and machine learning, with an emphasis on pushing beyond standard simulation limits toward methods that enable new regimes of scale, efficiency, and insight. I’m especially motivated by work that treats machine learning as a tool for scientific discovery and connects computation with clear, interpretable representations of complex systems, and I’m always open to research collaborations and ML-related internships.
