What happens when particles interact at strong coupling? At weak coupling, quantum field theory gave us a beautiful map of consistent theories: the gauge theories of the Standard Model. At strong coupling the map is almost blank, even though it contains the strong force and string theory. The difficulty is unitarity, a fully nonlinear equation that ties all processes together. Our group has developed a new way to explore this territory, reviving the 1960s S-matrix bootstrap with machine-learning tools, and recently built, for the first time, four-dimensional amplitudes that conserve probability exactly in two-particle collisions while including particle production. This internship uses these tools on a strikingly simple question: how fast can an interaction grow with energy? A universal limit, the Froissart bound, has been known since the 1960s, yet no one knows how real theories approach it. Protons look larger the harder they collide, a growth attributed to the mysterious Pomeron, and it is inside the Pomeron of early string models that the graviton was first discovered. The work can be analytic, numerical, or both. You will join a small, friendly team in Annecy, with regular one-to-one supervision and visits to CERN. No prior experience with machine learning is needed.