Software Architect: AI Systems & Co-Design
Join an early-stage, VC-backed startup working with leading frontier model labs to build the tooling that makes autonomous chip design possible. The team includes talent from NVIDIA, Groq, AMD, Intel, and MediaTek, and the company is growing fast with exponential revenue growth.
About the role:
- Define how complex AI workloads map onto high-performance silicon subsystems
- Map frontier model workloads to accelerator architectures and design hardware-software interfaces
- Optimise key kernels and explore how AI agents can accelerate architectural exploration and co-design workflows
- Investigate research questions at the edge of AI systems, ML-for-EDA, memory optimisation, and accelerator architecture
- Work directly with the founders and collaborate with frontier model researchers to shape technical direction
What we're looking for:
- Strong experience in AI systems, accelerator software, or hardware-software co-design
- Experience mapping and optimising AI workloads on GPUs, custom accelerators, NPUs, Tensor cores, or NVDLA-based systems
- Strong software engineering background and ability to build reliable systems
- Understanding of performance bottlenecks in AI infrastructure: memory bandwidth, compute density, SRAM capacity, and PPA tradeoffs
- Very strong communication skills
Nice to have:
- Experience with kernel optimisation, memory management, quantisation, or sparse tensor compression
- AI research, ML-for-EDA, or agentic systems experience
