It started with pybind11 envy. Back in 2020, I used pybind11 to create a Python extension for my company's C++ pattern discovery engine. It took only a few days but enabled our data scientists to make breakthrough discoveries in cancer research. There is something magical when compiled code comes alive in a REPL like IRB.

I wanted that experience in Ruby, but the tooling was not there. Thus began a 5-year journey to build a modern Ruby ↔ C++ toolchain. The result now powers Ruby bindings for libraries such as PyTorch, faiss and OpenCV.

This talk will walk through the toolchain:

• ffi-clang: parse C++ headers using libclang
• ruby-bindgen: generate bindings from the clang AST
• Rice: Ruby <-> C++ interoperability
• CMake: cross-platform native builds

We will cover challenges like ownership, garbage collection, templates and method overloading. Attendees will leave knowing how to bridge Ruby and modern C++ for both small utilities and large, complex libraries.