All my public codes are hosted on github.

A selection of the more useful ones:

• hmpdf: Computes PDFs of tSZ and lensing fields within the halo model. An extension allows halo-model calculations of (r)tSZ spectral distortion monopoles.
• voxelize: Maps particle catalogs from cosmological simulations to regular grids. Optional GPU acceleration.
• group_particles: A solution to the common problem of finding all simulation particles in the vicinity of halos and computing summary statistics on them. The code are header-only C++ templates, allowing for great flexibility with no runtime penalty.
• capped_mean: Small pytorch extension for efficient batch-wise averages where individual batch elements have varying data size. This is a common operation in set-based architectures. A CUDA version allows seamless GPU utilization.

I am proficient in C, C++, and Python. I have some knowledge in CUDA, Mathematica, and MySQL.