Small math and physics experiments, simulations, and visualizations that I did because curiosity or fun. This repo is a playground for testing ideas: stochastic processes, numerical integration, optics, sampling, quantum-inspired numerics, and a few supporting math utilities.

• src/brownian/ Brownian motion experiments in 1D and 3D, plus scripts for generating animations.
• src/mh/ Metropolis-Hastings sampling with a Gaussian target and a lightweight sampler implementation in PyTorch.

• src/integrals/line_integral.py Numerical line integral along a parametric curve.
• src/integrals/riemann.py Riemann-sum and trapezoidal integration experiments, with derivative and cumulative-integral plots.
• src/integrals/surface_integral.py More structured numerical surface-integral code for scalar surface integrals and flux through parametric surfaces.

• src/optics/refraction.py Refraction at a planar interface via Snell's law, including static plots and GIF animation.
• src/optics/ Also contains rendered outputs such as refraction.png and tir.gif.

• src/dft/dft.py A finite-difference density-functional toy model for hydrogen-like radial states.

• src/physical/spherical_armonics/main.py Legendre and associated Legendre polynomial experiments connected to spherical harmonics.
• src/physical/magnetic_field/main.py Early magnetic-field utilities based on point or wire geometry.
• src/experimental_magnetic_field/main.py Simple plotting of experimental magnetic-field versus current data.

• src/descompositions/main.py Gram-Schmidt orthonormalization, QR-style reconstruction, and reconstruction error checks.

• src/kans/main.py A scratch implementation of a Kolmogorov-Arnold Network style layer trained on a toy regression task.

• src/matplotlib/main.py Plotting helpers for heat-equation style comparisons, errors, scalar fields, and 3D views.

• src/async/async.py Small async vs threading comparison for I/O-bound tasks.

• src/a.ipynb Notebook workspace for ad hoc exploration.

Some experiments use extra scientific or ML dependencies such as numpy, matplotlib, scipy, and torch as usual, so get fun!