I'm the co-author.The code is in a very very bad state right now, but the architecture is pretty ok to explain. In step 3, we translate from the Python frontend

I'm the co-author.The code is in a very very bad state right now, but the architecture is pretty ok to explain. In step 3, we translate from the Python frontend to the LLVM IR. In step 4 we compile it down to an object file using the LLVM backend `llc`. This object file gets loaded into the kernel and it is what actually contains the eBPF bytecode.