Principal Software Architect

We’re supporting a well-funded semiconductor company developing a new programmable architecture for AI and high-performance computing.

They’re hiring a senior hands‑on architect who can connect accelerator architecture, systems software and data‑centre workloads. This is broader than a traditional compiler or GPU‑kernel role: you’ll help define how complete workloads execute from frameworks through compilers and runtimes to clusters of accelerators.

What you’ll work on

• Architect the end-to-end execution stack for AI and compute workloads
• Present at conferences and engage with customers
• Translate data‑centre workload requirements into hardware and software decisions
• Shape framework, compiler, runtime, driver and programming‑model interfaces
• Analyze compute, memory, communication and scheduling bottlenecks
• Define workload mapping across different processing engines
• Evaluate multi‑accelerator execution, scale‑up and scale‑out behavior
• Prototype critical ideas and guide teams through implementation

What we’re looking for

• Principal, Distinguished or equivalent architecture‑level experience
• Strong understanding of GPU, accelerator or heterogeneous‑compute systems
• Depth across several layers of the compute stack—not only one compiler component
• Experience with production or data‑centre‑scale AI/HPC workloads
• Knowledge of parallel programming, memory systems and workload execution
• Ability to remain technically hands‑on in C++ or comparable systems software

Particularly relevant backgrounds include GPU architecture, AI systems, accelerator software, HPC, ML compilers, distributed compute and hardware/software co‑design.