I was recently contracted out to find the distribution of some experimental data and then find a way to generate random numbers from that distribution using an RNG that can only generate random numbers uniformly on the [0, 1) interval.

An interesting project, led me down a bit of a rabbit hole of probability density function transformation techniques. I eventually settled on a rejection-selection technique outlined in a 2011 article from the Journal of Statistical Physics by Nader Mohamed to generate Maxwell random variables, and an extension of the Box-Muller algorithm to generate log-normal random variables.