I understand that separation of the grammar and semantics has its benefits. You can use the same grammar description with different semantic rules sets etc. It's, indeed, a clean and interesting approach.
But, as I understand, Ohm still has no support for context-sensitive grammars, which is more important to have in many cases, than proper left recursion handling.
And oMeta had another nice feature: meta-rules (higher-order rules) which is absent in Ohm, if I understand correctly.
Ohm tries to be very user-friendly, but at the price of droping the functionality. So in this case Ohm is not a modern replacement for oMeta (which had the AST transforming features -- very important for making compilers!).
I'm not trying to be negative and I really wish a big success to your team!