LLM Inference Performance & Evals Engineer

<div class="content-intro"><p><span data-contrast="none">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. </span><span data-ccp-props="{"134233117":false,"134233118":false,"201341983":0,"335559685":0,"335559737":240,"335559738":240,"335559739":240,"335559740":279}"> </span></p> <p>Cerebras' current customers include top model labs, global enterprises, and cutting-edge AI-native startups. <a href="https://openai.com/index/cerebras-partnership/">OpenAI recently announced a multi-year partnership with Cerebras</a>, to deploy 750 megawatts of scale, transforming key workloads with ultra high-speed inference. </p> <p>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.</p></div><h4>About The Role</h4> <p>Join the inference model team dedicated to bring up the state-of-the-art models, numerically validating and accelerating new model ideas on wafer-scale hardware. You will prototype architectural tweaks, build performance-eval pipelines, and turn hard numbers into changes that land in production.</p> <h4>Key Responsibilities</h4> <ul> <li>Prototype and benchmark cutting-edge ideas: new attentions, MoE, speculative decoding, and many more innovations as they emerge. </li> <li>Develop agent-driven automation that designs experiments, schedules runs, triages regressions, and drafts pull-requests. </li> <li>Work closely with compiler, runtime, and silicon teams: unique opportunity to experience the full stack of software/hardware innovation. </li> <li>Keep pace with the latest open- and closed-source models; run them first on wafer scale to expose new optimization opportunities. </li> </ul> <h4>Skills And Qualifications </h4> <ul> <li>3 + years building high-performance ML or systems software. </li> <li>Solid grounding in Transformer math—attention scaling, KV-cache, quantisation—or clear evidence you learn this material rapidly. </li> <li>Comfort navigating the full AI toolchain: Python modeling code, compiler IRs, performance profiling, etc. </li> <li>Strong debugging skills across performance, numerical accuracy, and runtime integration. </li> <li>Prior experience in modeling, compilers or crafting benchmarks or performance studies; not just black-box QA tests. </li> <li>Strong passion to leverage AI agents or workflow orchestration tools to boost personal productivity.</li> </ul> <h4>Assets</h4> <ul> <li>Hands-on with flash-attention, Triton kernels, linear-attention, or sparsity research.</li> <li>Performance-tuning experience on custom silicon, GPUs, or FPGAs. </li> <li>Proficiency in C/C++ programming and experience with low-level optimization. </li> <li>Proven experience in compiler development, particularly with LLVM and/or MLIR. </li> <li>Publications, repos, or blog posts dissecting model speed-ups. </li> <li>Contributions to open-source agent frameworks.</li> </ul><div class="content-conclusion"><h4><strong>Why Join Cerebras</strong></h4> <p>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 join ... (truncated, view full listing at source)