We study flavour violation in a supersymmetric SO(10) implementation of the type II seesaw mechanism, which provides a predictive realization of triplet leptogenesis. The experimental upper bounds on lepton flavour violating processes have a significant impact on the leptogenesis dynamics, in particular they exclude the strong washout regime. Requiring successful leptogenesis then constrains the otherwise largely unknown overall size of flavour-violating observables, thus yielding testable predictions. In particular, the branching ratio for μ → eγ lies within the reach of the MEG experiment if the superpartner spectrum is accessible at the LHC, and the supersymmetric contribution to ε_K can account for a significant part of the experimental value. We show that this scenario can be realized in a consistent SO(10) model achieving gauge symmetry breaking and doublet-triplet splitting in agreement with the proton decay bounds, improving on the MSSM prediction for α₃(m_Z), and reproducing the measured quark and lepton masses.