Principal ML Investigator

Cerebras Systems•Sunnyvale, CA

19h

About The Position

Cerebras Systems is seeking a principal investigator to lead a new Machine Learning (ML) effort that aligns with existing teams. This role involves partnering with ML leaders to formulate the new effort, build a team, and develop new capabilities. The new team will coordinate with existing ML teams: Field ML (customer-facing), Applied ML (building new ML capabilities), and Core ML (adapting ML algorithms for Cerebras hardware). The new team's focus areas may include post-training and reinforcement learning, dataset curation and optimization, LLM pretraining, sparsity techniques, and specific domains like coding agents, reasoning agents, and generative language, image, or video.

Requirements

PhD in Computer Science or related field.
Strong grasp of ML theory in one or more of the specified areas (post-training/RL, dataset curation, LLM pretraining, sparsity, coding/reasoning/generative agents).
Proven experience engineering ML systems for scale or production deployment.
Experience leading a team of researchers or engineers.

Nice To Haves

Track record of patents or publications in top-tier conferences or journals.
Experience with large language models (e.g., GPT family, Llama).
Experience with distributed training concepts and frameworks.
Experience in training speed optimizations, such as model architecture transformations to target hardware, or low-level kernel development (e.g., Triton).
Ability to analytically model or optimize system performance.

Responsibilities

Build up a team capable of industry research and advanced development.
Organize various advanced development topics into a cohesive agenda.
Adapt novel algorithms and model architectures to run on the Cerebras platform.
Systematically train, tune, and evaluate models to guide/advise production scenarios.
Collaborate with other teams to co-design next-generation hardware and software architectures.
Collaborate with external partners (customers, academic) to drive insight and credibility.