Rainforest is conceived as a large-scale self-regulating audiovisual ecosystem; its morphological and behavioral evolution over time blends internal motivations and external pressure from exposure to the external world. In other words, Rainforest interfaces spontaneous natural behavior in the tangible world of everyday life with artificial life in a parallel synthetic universe. Rainforest consists of simple, locally interacting particle objects. However, particles coalesce into larger, more complex structures, complexity that echoes in the emergent sound produced by the installation. Rainforest features a machinelearning component; it aims to optimize the aesthetic experience by coordinating artificial life in relation to audience response captured by computer vision.