Architecture · Rust · C++ · Python
One core system for research and production
A cross-language architecture for a real-time vision system, designed to keep research and production work connected to the same core implementation.
Challenge
Critical system logic had to serve an established C++ production library as well as fast-moving research and prototyping workflows. Separate implementations would have created duplicated logic and different sources of truth.
Contribution
Founded the C++/CUDA-to-Python binding layer over the original C++ paths — the initial sketch of what became the client's current system, still used throughout for research and experiments — and now delivers the Rust reconstruction core back into that system as a dependency-free C-ABI library — a binary interface callable from any language that speaks C's conventions — with a versioned compatibility handshake, a panic-safe boundary, a documented C header, and three-OS CI.
Outcome
One verifiable core serving production C++ and research Python alike — the system's shape was sketched through the bindings, and the Rust engine now plugs into it across a stable C ABI.
Client work is described at a technical, client-agnostic level.