High Performance Compute Responsible Engineer

You'll own the compute hardware from architecture through bring up — there's no hand-off to another team mid-stream. You pick the components, you design the board, you're the one powering it on and debugging it with the software team

• Design the compute boards that deliver ≥10 TFLOPS of processing power for onboard AI/ML inference and science data processing within the Relay Data Center, owning the hardware from schematic capture through board-level validation
• Architect and implement power delivery networks for high-complexity carrier boards, including 500W+ multi-rail designs with tight sequencing integrating GPU or CPU compute modules onto custom PCBs with high-speed coherent interconnects