Predictive science is a multilevel process requiring observational data and a model/hypothesis which will have to calibrated and validated to eventually the predict the quantity of interest. For decades, the method of choice for model calibration has been the Monte Carlo simulations. Here we present and discuss serial and parallel implementations of a collection of the popular powerful Monte Carlo techniques that can aid inference and uncertainty quantification in Machine learning and Bayesian problems. Emphasis in the development of this open-source package, named ParaMonte, has been on user-friendliness, accessibility from different programming languages and platforms, high-performance, parallelism and scalability, as well as reproductivity and comprehensive post-processing and visualization of the simulation results. The MIT-licensed ParaMonte library is accessible from Python, MATLAB, R, Julia, Fortran, C++, and C, and is permanently located at https://github.com/cdslaborg/paramonte