Senior System Software Engineer (AI Runtime & Compute)
How much AI performance is lost between what the compiler produces and what the hardware actually executes?
I’m working with a venture-backed AI infrastructure startup tackling that problem through a new compiler and runtime platform. The aim is to help developers get more from different accelerators while reducing the engineering work involved in moving between hardware platforms.
The team brings together deep experience in compilers, runtimes and processor architecture. They’re hiring a Runtime Engineer to build a central part of the execution stack.
The opportunity
You’ll work where compiler output meets real hardware: deciding how work is scheduled, how parallelism is used and where execution loses performance.
Your work will include:
- Developing runtime components that support multiple hardware targets.
- Exploring scheduling and workload partitioning approaches.
- Building experiments that reveal where execution stalls and why.
- Measuring generated code on hardware and feeding findings into the compiler team.
- Turning promising ideas into reliable, maintainable systems software.
What would make you a strong fit
- Strong C++ experience and a background in runtime or compiler systems.
- A solid grasp of concurrency, asynchronous programming and memory hierarchies.
- Knowledge of operating-system kernel or hypervisor development.
- An interest in investigating performance through code, measurements and hardware behaviour.
GPU programming, CUDA/ROCm, HPC or experience with large compute systems would also be valuable.
This is an opportunity to influence core execution technology alongside experienced systems engineers, with equity participation and a direct connection between your engineering decisions and platform performance.
