Magnetic instability in metals often reflects special features of their Fermi surfaces, e.g., so called nesting. The evolution of the Fermi surface as a function of external parameters provides insight into the mechanism of magnetic instability. The goal of this work is, in the first step, to use band structures or tight binding models calculated for various parameters such as doping or strength of the electron-electron interaction and create interactive 3D visualizations of the corresponding Fermi surfaces. In the second step, machine learning will be  used to develop models describing the Fermi surface as a function of smoothly varying parameters. The target material will be RuO2. Application to other selected materials is also possible.