Turing.jl

Why we included this project

Turing.jl is a probabilistic programming library for Julia that lets you write a model in code close to how you would write it on paper, then runs Bayesian inference over it. Instead of a single fitted value, you get a posterior distribution for each parameter, which is what you need when decisions rest on calibrated uncertainty rather than a point estimate. You declare priors and likelihoods inside an @model block, draw samples with NUTS, HMC, or other MCMC samplers, and switch to variational inference when runtime matters. That combination makes it a natural choice for hierarchical models, A/B tests, and time series work in Julia, all backed by proper documentation and an ecosystem maintained by researchers.