Abstract: Understanding the properties of atomic nuclei and their connection to astrophysical phenomena requires reliable theoretical models together with increasingly precise experimental constraints. In this seminar, I will present some of my recent work spanning nuclear structure, nuclear astrophysics, and data-driven nuclear modelling. I will first discuss the electric dipole response of finite nuclei and how dipole polarizability can provide constraints on the density dependence of the nuclear symmetry energy and on neutron-star properties. I will then address low-energy electric dipole strength in neutron-rich nuclei and its impact on astrophysical neutron-capture and photodisintegration reaction rates, with particular emphasis on the role of strength near the particle-emission threshold. Finally, I will present recent applications of machine-learning and quantum-machine-learning methods to nuclear mass predictions, including ensemble-based approaches and physics-informed feature representations. These studies illustrate complementary ways in which nuclear observables, microscopic models, and modern computational methods can be combined to improve the predictive description of nuclei across the nuclear chart.