Neutrino flavor oscillation in compact object mergers will significantly affect the merger dynamics and the electron fraction. In particular, Fast Flavor Instability close to the central object in neutron star merger simulations can change r-process abundance. Over the years, many approaches have been taken to include flavor oscillation during post-processing, while a couple of recent works include them in situ. However, given how expensive numerical transport is, solving the quantum kinetic equation directly in a merger simulation is far from achievable. In this talk, I will present our efforts to incorporate Fast Flavor Oscillation (FFO) in mergers using subgrid modeling. I will present a machine-learning model and an analytical mixing scheme that can predict the outcome of FFO on post-merger snapshots and compare their performance. I will also discuss the challenges of incorporating these models in the available neutrino transport scheme.