Efficient sampling is fundamental to modern statistical inference—from Bayesian modeling and posterior estimation to uncertainty quantification and probabilistic AI. Yet sampling from high-dimensional constrained distributions remains a major computational bottleneck.

This talk introduces volesti, an open-source C++ library with R and Python interfaces, designed to make sampling and numerical integration over convex domains both scalable and practical. It implements advanced geometric random walks and multiphase Monte Carlo algorithms that scale to thousands of dimensions.

We’ll explore:

1. Why high-dimensional sampling is critical across scientific and engineering domains
2. How volesti tackles problems that standard samplers can’t handle
3. Use cases in Bayesian statistics, finance, and computational biology

Whether you're an engineer, data scientist, open-source contributor, or just curious about the geometric side of statistical inference, this talk connects algorithms with practical, open-source tooling.