The PhD proposal investigates the nucleon’s three-dimensional structure using Generalized Parton Distributions (GPDs). GPDs give access to the spatial distribution of quarks and gluons, the energy-momentum tensor, and thus information on spin, internal pressure, and mass. Two main challenges arise: scarce exclusive experimental data and the high cost of precise lattice-QCD simulated observables. The project comprises two parts: (I) generate new lattice-QCD simulations of GPD moments, improve algorithms, and perform continuum extrapolations; (II) create machine-learning tools to tackle the ill-posed inverse problem and conduct global fits that combine experimental and simulated data. The work will be carried out at the European Joint Virtual Lab AIDAS shared between Julich Forschungszentrum (Germany) and CEA (France), with equal time spent in each country. Required skills include quantum field theory, object-oriented programming (C++, Python), and high-performance computing. The ultimate goal is the first reliable extraction of the nucleon’s 3-D structure, informing future facilities such as the EIC and EicC.