Cerebras Systems designs wafer-scale AI accelerator chips, with its CS-2 and CS-3 processors each containing hundreds of thousands of cores on a single silicon wafer, fundamentally rethinking the architecture of AI compute hardware. The company's engineering challenges span the full hardware-software stack: designing chips that manage thermal dissipation and yield across an entire 300mm wafer, building compilers that map neural network operations onto massively parallel core arrays, and developing software frameworks that make wafer-scale computing accessible to ML researchers. Their inference offering competes on latency and throughput with GPU-based solutions, requiring optimization across kernel-level scheduling, memory hierarchy management, and model parallelism strategies. Cerebras' hiring patterns signal the AI hardware industry's push beyond GPU-centric architectures, with demand for engineers who can work across chip design, systems software, and ML compiler optimization.
| Location | Listings |
|---|