Software
Open-source software projects and libraries.
Featured Projects
volesti
C++ library (with R bindings) for high-dimensional sampling and volume computation on convex bodies. Implements geometric random walks (Hit-and-Run, Billiard Walk, Hamiltonian Monte Carlo), volume estimation algorithms, and uniform/Gaussian sampling. Released on CRAN as the R package volesti.
Boost.Geometry
Generic C++ geometry library, part of the Boost C++ Libraries collection. Provides geometric algorithms (distance, intersection, union, convex hull, etc.) for Cartesian and geographic coordinate systems. Core developer since 2016 — contributed geodesic algorithms, robust predicates, and spatial index improvements.
Dingo
Python package for sampling the flux space of metabolic networks using geometric random walks. Provides fast MCMC methods to sample steady-state flux distributions in genome-scale metabolic models, enabling statistical analysis of metabolic phenotypes.
Volume Approximation
Open-source C++ software for practical volume approximation of convex polytopes in high dimensions. Implements randomized algorithms with theoretical guarantees, using geometric random walks and simulated annealing. Applied to financial portfolio analysis and metabolic network modeling.
ResPol
Software package for computing projections of Newton polytopes of the sparse resultant. Implements oracle-based output-sensitive algorithms for high-dimensional polytope computation. Used in computational algebraic geometry and elimination theory.
HeaDDaCHe
Hashed Dynamic Determinant Computation for faster geometric algorithms. Implements a caching scheme for sequences of determinantal predicates used in convex hull computation, point location, and other geometric primitives, achieving significant speedups over direct determinant evaluation.