Given the spectrum of an exoplanetary atmosphere, a multi-parameter space is swept through to find the best-fit model. Known as atmospheric retrieval, such methods are time-consuming and thus there is a compromise between physical complexity of the model and computational feasibility. I will present our adaptation of the random forest method, trained on a precomputed grid of atmospheric models, which retrieves posterior distributions of the model parameters. The use of a precomputed grid allows a large part of the computational burden to be shifted offline. The forest can estimate the sensitivity of the measured spectrum to the model parameters, and quantify the information content of the data.