Applied AI ML Scientist
Cerebras Systems, Inc.Visit Cerebras Systems, Inc. website
Cerebras builds wafer-scale AI computing systems and a cloud inference platform for training, fine-tuning, and serving AI models.
Sunnyvale, California, United States
About Cerebras Systems, Inc.
Cerebras Systems is an AI-infrastructure company founded in 2015. It sells rack-scale wafer-scale computing systems and provides cloud-based, API-accessible AI inference alongside on-premises deployments.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will scope customer AI engagements and develop custom large-scale models and agentic systems. You will design training recipes, preprocess data, fine-tune and align models, scale workloads across clusters, analyze training behavior, and translate customer needs into technical solutions.
Requirements
- Machine learning
- deep learning
- transformer
- mixture of experts
- multimodal model
- sequence model
- scaling law
- training dynamics
- large-model training
- model fine-tuning
- Python
- PyTorch
- distributed training
- distributed data processing
- data curation
- communication
Responsibilities
- Identify AI approaches for customer business problems
- Scope engagements through feasibility and data-readiness assessments
- Define project milestones, success metrics, and evaluation benchmarks
- Architect and execute training recipes for custom models
- Implement continuous pre-training, supervised fine-tuning, RLHF, and DPO strategies
- Own training pipelines from data preprocessing and tokenization through hyperparameter tuning and loss analysis
- Analyze model convergence, loss dynamics, and gradient stability
- Scale multi-billion-parameter training workloads across clusters
- Build components for agentic systems including tool use, long-context reasoning, and multi-step planning
- Translate customer requirements into training recipes
- Share customer feedback with research and engineering teams
- Create internal playbooks from successful customer projects