C++ / Python Integration Engineer

Ensigncode is hiring a C++/Python Integration Engineer to own API and developer experience on our GPU optimization team: building clean, stable C++ APIs for GPU functionality and exposing them to Python via PyBind11 and PyTorch extensions. CUDA experience is valuable here, but deep kernel expertise is not required.

The best kernel in the world is only useful if someone else on the team, or a client, can actually call it reliably. As C++/Python Integration Engineer, you’ll own that layer: designing clean, stable C++ APIs for GPU functionality and exposing them to Python through PyBind11 and PyTorch extensions, with careful attention to tensor and device compatibility, memory ownership, and error handling.

You don’t need to be a kernel specialist for this role, but you do need real depth in modern C++, Python, and build tooling like CMake, plus the API design instincts that make optimized GPU code something other engineers can pick up without reading the kernel source first.

Responsibilities

• Build clean, stable C++ APIs for GPU functionality and expose them to Python using PyBind11 and/or PyTorch extensions
• Manage tensor/device compatibility, memory ownership, error handling, build systems, packaging, and version compatibility
• Create tests and examples that make optimized functionality straightforward for other engineers to use
• Support profiling and debugging across the C++/Python boundary

Core requirements

• Modern C++; Python; PyBind11; PyTorch extension development; build tools such as CMake; API design and testing
• CUDA experience is valuable, but deep kernel expertise is not required

Team contribution

Makes the team's GPU work usable, testable, and maintainable by the wider software stack.

Works most closely with