Design optimisation is a growing area of interest for many industrial fields. With requirements con- tinually being pushed, economic and time constraints faced by engineers is constantly being stretched [1]. We present a framework for optimising the performance of existing designs for which a discrete un- structured mesh is available. The geometry is parameterised and the design space is explored using Bayesian methods. The geometry is modified using Radial Basis Function (RBF) interpolation to morph the baseline design. The computation cost and wall clock time can be balanced depending on the users requirements by modifying the quantity of training data and number of design iterations to perform. The methodology is applied to an industry sponsored test case to minimise the drag within the transonic regime of a spaceplane currently under development [2]. The lift and drag coefficients for the baseline design is calculated using CFD at Mach 1.2 and incidence 4◦. The lift coefficient for morphed designs is constrained by changing the incidence for the morphed designs. The total volume of the morphed designs is also constrained to that of the baseline design by modifying the fuselage length after all independent parameters have been applied. The framework is tested with different quantities of training data for a maximum of 20 total objective function evaluations. In all cases, a reduction in drag coefficient of between 1 − 2% was achieved. We then compare the variations of computational and wall clock time compared to the variation in training data quantity.