Cerebras Systems builds the world's largest AI chip, 56 times larger than GPUs. Our novel wafer-scale architecture provides the AI compute power of dozens of GPUs on a single chip, with the programming simplicity of a single device. This approach allows Cerebras to deliver industry-leading training and inference speeds and empowers machine learning users to effortlessly run large-scale ML applications, without the hassle of managing hundreds of GPUs or TPUs.
Cerebras' current customers include top model labs, global enterprises, and cutting-edge AI-native startups. OpenAI recently announced a multi-year partnership with Cerebras, to deploy 750 megawatts of scale, transforming key workloads with ultra high-speed inference.
Thanks to the groundbreaking wafer-scale architecture, Cerebras Inference offers the fastest Generative AI inference solution in the world, over 10 times faster than GPU-based hyperscale cloud inference services. This order of magnitude increase in speed is transforming the user experience of AI applications, unlocking real-time iteration and increasing intelligence via additional agentic computation.
Cerebras builds wafer-scale AI processors—single chips delivering tens of PB/s of memory bandwidth and a dataflow architecture that accelerates at a granularity no multi-device system can match. The Advanced Technology Group (ATG) is Cerebras’ pathfinding organization. We work ahead of product to explore new architectures, demonstrate breakthrough performance on scientific and AI workloads, and shape the technical roadmap for future Cerebras hardware and software. Our work regularly appears at top-tier venues (Supercomputing, SIAM, IEEE, and NeurIPS) and directly influences the design of next-generation wafer-scale systems.
We are seeking Compiler Engineers to join a small team of specialists working on our emerging Tungsten language compiler. Tungsten is Cerebras’ dataflow programming language, purpose-built for wafer-scale hardware. You will work on the Tungsten compiler from language design through code generation, building the toolchain that translates high-level intent into efficient execution across hundreds of thousands of cores with a memory and interconnect model unlike anything in conventional computing.
This is not incremental work on an existing backend. The architecture is new, the programming model is new, and the compiler is where those two things meet. You will collaborate closely with Cerebras’ ASIC, kernel, and AI teams, and your design decisions will directly shape both the language and the hardware it targets. Beyond the compiler itself, the broader toolchain—runtime, debugger, simulator—is still being built, and we are equally interested in engineers who want to own those pieces of the developer experience on novel hardware.
- Design and implement compiler passes across the Tungsten toolchain: mid-end optimization, backend code generation, instruction scheduling, register allocation, assembler, and linker.
- Co-design language constructs that improve expressiveness and performance for dataflow execution on wafer-scale hardware.
- Develop and iterate on code generation strategies for complex scientific and AI workloads, analyzing performance bottlenecks and closing the gap between peak and achieved throughput.
- Extend the compiler to support future hardware architectures as they move from design to silicon.
- Work directly with ASIC architects and application researchers to inform hardware-software co-design decisions.
- PhD in Computer Science or Computer Engineering preferred; exceptional candidates without a graduate degree who demonstrate equivalent depth through published research, significant open-source contributions, or a strong industry track record are encouraged to apply.
- Substantial experience in compiler development: IR design, optimization passes, code generation, or backend implementation for novel or non-standard architectures.
- Strong grasp of computer architecture: instruction sets, memory models, dataflow execution, and how hardware constraints shape compilation strategy.
- Systems-level programming ability in C; comfort reasoning about performance at the instruction and memory-access level.
- Ability to think about compilation as a design problem, not just an implementation task: you should have opinions about how language semantics, compiler IR, and hardware capabilities interact.
- Excellent communication and interpersonal skills: able to work effectively in a small, fast-moving team where compiler, architecture, and application concerns are deeply intertwined.
- Experience with compilers for spatial, dataflow, or CGRA architectures where the compilation model diverges significantly from conventional CPU/GPU targets.
- Exposure to ML compiler frameworks (MLIR, XLA, TVM) and understanding of how AI workloads map to hardware.
- Experience with multi-dimensional data representations, tiling strategies, and vectorized operations.
- Track record of published research or patents in compilers, programming languages, or architecture.
- Experience building runtime systems, debuggers, or architecture simulators, particularly for non-standard hardware.
- Understanding of parallel/distributed systems and cluster computing.
- Build a compiler for hardware that doesn’t exist anywhere else. The architecture is the constraint and the opportunity.
- Publish and open-source your research. We present at Supercomputing, SIAM, IEEE, NeurIPS, and beyond.
- Work on the fastest AI system in the world, with direct access to the hardware your compiler targets.
- Join at a pivotal moment: Cerebras is pre-IPO with strong commercial traction and rapid growth.
- Be part of a small, technical team with high autonomy, minimal bureaucracy, and a culture that values depth over hierarchy.
We are hiring for multiple positions across experience levels. If this work resonates, we encourage you to apply.
People who are serious about software make their own hardware. At Cerebras we have built a breakthrough architecture that is unlocking new opportunities for the AI industry. With dozens of model releases and rapid growth, we’ve reached an inflection point in our business. Members of our team tell us there are five main reasons they joined Cerebras:
- Build a breakthrough AI platform beyond the constraints of the GPU.
- Publish and open source their cutting-edge AI research.
- Work on one of the fastest AI supercomputers in the world.
- Enjoy job stability with startup vitality.
- Our simple, non-corporate work culture that respects individual beliefs.
