Estimate a high dimensional SDE model

hi there, I have a high dimension (n=6) Euler-Maruyama model to estimate. The state variables are mostly latent except for 2. Is there an example that you could point me to so I can see how find_MAP or other optimizers can handle nested functions in a SDE model? many thanks, am a first time user of pymc3, I am not clear in terms how to specify the user-defined drift and diffusion functions with empirical observations.