Most CS undergraduate mathematics stops just short of proof-based courses, which is a shame. A proof-based linear algebra course would be the next course that

Most CS undergraduate mathematics stops just short of proof-based courses, which is a shame. A proof-based linear algebra course would be the next course that is typical in the progression. If you've already done that, then an introductory real analysis course. If you want something a bit more practical then maybe a probabilistic modeling course? ( think Gaussian mixture models, bayesian networks, plate models).