You can use multiple distributed rendering systems together with Smedge if you are careful. The thing to watch out for is to make sure that Engines that are part of the distributed rendering pool for the other product, for example, those listed in the rayhosts file for mental ray renders, are not also enabled to take on additional work assigned by Smedge.
You can always disable your Engines manually before submitting a Job that will use them via an alternate distributed rendering system. But if you have other work queued, you may not want to sit there waiting for machines to finish what they were working on before you submit your render. In this case, you can create a Generic Script Job that will disable all of the engines you need, and attach a command line to the "Job Finished" event to then submit your mental ray render Job.
Use the Engine command line Shell in the Generic Script Job to disable the Engine from doing other work. Group the Engines into a Pool, and use the Pool for the disabling Generic Script Job. For the mental ray rendering Job, make sure to assign the Job to a specific single machine. That machine should be the mental ray "master" and should be set up with the appropriate configuration files for that machine to distribute work to the other machines. Note that some Products, for example: finalRender for Maya, allow you to configure the distributed hosts as part of the command line, and thus as part of the Smedge Job information.