At BRITE we combine strong backgrounds in fluid mechanics and system identification. Building on that foundation, we have developed nonlinear data-driven modelling tools that are suited to model challenging thermal and fluid-dynamics problems, involving nonlinearity, unsteadiness, non-stationarity, and turbulence. These modelling techniques can be widely applied, also beyond thermal and fluid mechanics. In combustion, these approaches have been used to tackle the main challenges related to reacting flow problems, namely the curse of dimensionality and the complex nonlinear turbulence-chemistry interactions (e.g., sub-grid closures).