carma_pack is an MCMC sampler for performing Bayesian inference on continuous time autoregressive moving average models. These models may be used to model time series with irregular sampling. The MCMC sampler utilizes an adaptive Metropolis algorithm combined with parallel tempering.