This presentation will introduce the development of a data-driven coupled-cluster (DDCC) approach that integrates quantum machine learning (QML) with molecular electronic structure calculations. Our approach is built around a quantum–centric framework where quantum computing resources are utilized to predict coupled-cluster excitation amplitudes. These amplitudes are conventionally expensive to obtain through iterative coupled-cluster solvers and are central to the accurate computation of ground state molecular energies. In our work, wavefunction derived descriptors are used as input features to QML models, such as parametrized quantum circuits and quantum kernel based models, either allowing a full bypass of solving the amplitude equations or enabling the use of QML predicted excitation amplitudes as warm starts for the iterative solvers. In this presentation, I will discuss our investigations into the learning of parametrized quantum circuits and quantum kernels for the task of accurately predicting these excitation amplitudes, along with results on the transferability of the QML models assessed by their ability to predict excitation amplitudes for out-of-distribution molecular systems. These results will highlight the potential of QML models in the broader context of a hybrid quantum–classical pipeline for quantum chemistry.