Applied AI/ML Scientist

Cerebras Systems • Sunnyvale, CA

1d

About The Position

Cerebras Systems is seeking an Applied AI Scientist to join its FieldML team. The role focuses on developing and customizing large language models and other large-scale deep learning models to solve specific customer problems. The scientist will build and train models, bridging state-of-the-art research and practical applications by helping customers use the Cerebras Wafer-Scale Engine for AI initiatives.

The work spans training bespoke models from scratch, fine-tuning and optimizing large language models for industry-specific needs, and designing components for custom agentic systems. The role also involves working directly with customer stakeholders, translating requirements into technical plans and training recipes, and collaborating with internal research, engineering, ML, and product teams to improve the software stack, hardware utilization, and delivery practices.

Requirements

• Master’s or PhD in Computer Science, Machine Learning, or a related field.
• Expert-level understanding of modern deep learning architectures, including dense transformers, mixture-of-experts models, multimodal models, and sequence models, as well as scaling laws and training dynamics.
• Proven experience training and/or fine-tuning large models with 1B+ parameters and direct experience with large-scale model-training challenges.
• Mastery of Python and PyTorch.
• Experience with distributed training frameworks and large-scale distributed data-processing pipelines and tools.
• Strong collaborative and communication skills, including the ability to work autonomously and in a team in a dynamic environment.
• Ability to manage multiple projects and adapt as customer needs evolve.
• Ability to present complex technical results to audiences ranging from C-level executives to research scientists.
• Ability to work collaboratively to solve customers’ unique challenges.

Responsibilities

• Collaborate with customer stakeholders to identify effective AI approaches for their business problems.
• Contribute to technical engagement scoping, including feasibility analysis, assessments of data quality, availability, and readiness, and model architecture selection.
• Define project milestones, success metrics, and evaluation benchmarks.
• Design and execute end-to-end training recipes for custom models, adapting architectures and recipes to customer performance and accuracy requirements.
• Develop adaptation strategies such as continuous pre-training on private datasets, supervised fine-tuning, and post-training alignment using RLHF or DPO.
• Own training pipelines, including data preprocessing, tokenization, hyperparameter tuning, and loss-curve analysis.
• Analyze model convergence on specialized hardware, including loss dynamics and gradient stability.
• Scale training workloads across Cerebras clusters and efficiently use hardware for multi-billion-parameter models.
• Build and optimize agentic system components focused on tool use, long-context reasoning, and multi-step planning.
• Serve as an AI/ML subject matter expert in technical discussions and translate customer requirements into training recipes.
• Build and maintain customer relationships as a trusted AI/ML expert.
• Represent customer needs to internal R&D and engineering teams to drive software-stack and hardware-utilization improvements.
• Partner with internal ML and product teams on model architecture prioritization, training recipes, and internal case studies.
• Turn successful customer projects into internal playbooks for the FieldML team.

Benefits

• Job stability with startup vitality.
• A work culture described as simple and non-corporate that respects individual beliefs.
• Opportunities to publish and open source cutting-edge AI research.
• Opportunity to work on one of the world’s fastest AI supercomputers.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior