Senior Product Manager, AI Models

About the role

Cerebras Systems builds the world's largest AI chip, delivering training and inference speeds over 10 times faster than GPU-based cloud services. As Product Manager for AI Models, you will lead the strategic model portfolio: deciding which models ship, how they perform and how they are discovered. You will partner directly with leading AI labs, drive launches that shape the industry and ensure every model on the platform delivers quality at unprecedented speed.

Key responsibilities

• Own the model roadmap: choose frontier and open-source models based on market demand, research trends and strategic fit, and build partnerships with model labs and open-source maintainers.
• Define quality standards across the model catalog, design benchmarks and evaluations, and close the feedback loop with customer insights.
• Help strategic customers integrate inference into their products and remove blockers.
• Lead model launches and create product marketing: demos, benchmarks, tutorials and documentation, plus technical content for developers and decision-makers.
• Select and prioritize performance optimizations (quantization, speculative decoding) and balance quality, latency, throughput and cost.
• Orchestrate launches across model enablement, optimization engineering, deployment, sales and marketing, and act as the voice of the customer to engineering and the voice of product to customers.

Required profile

• 5+ years as a product manager at or above Senior level.
• 5+ years of total technical work experience (e.g. software engineer, ML researcher, solution engineer).
• Ability to thrive in a fast-paced environment, with an entrepreneurial sense of ownership and project leadership.
• Knowledge of open-source models and generative AI research, and of the community ecosystem (PyTorch, Hugging Face, vLLM, SGLang).
• Comfortable using Python with the chat completions API for basic model testing.

Preferred requirements

• PM experience at a model training lab or at a company implementing open-source models.
• Solution engineering customer-facing experience, technical marketing and social media writing, and cross-functional project leadership.
• Experience writing model quality evaluations and system prompt harnesses, and application code for use cases such as code generation or deep research search.
• Expertise in agentic flows, LLM architectures, model compilers and optimization, and contributions to communities like vLLM, SGLang, PyTorch or Hugging Face.

Location: hybrid at the Sunnyvale, California or Toronto, Canada office, with remote work possible for candidates willing to travel 1-2 times per quarter.