I am not sure what type of post-processing you are doing and what the result you are looking for is, but I use NETLink extensively and it works very well in many situations. I do not have the code in front of me, but I do have a screenshot of an example where I plot a simple function. I can also extract the list of points in the graphics object if so desired.
I need a few details to know the best suggestion for you. This easiest thing to do is to interact with the kernel directly to send commands and retrieve the results (like I do with the plot here). You can also launch the front end,load your notebook and execute it. This has a few gotchas if I recall correctly, (it is tricky to know when the kernel is done evaluating). That may not matter to you. It would help to know what type of post-processing is going on and if you do anything with the results.