Amazon
Applied Scientist -- Foundation Models, SSO
$172K–$222K
Santa Clara, CA · Full-time · Posted Aug 17, 2026
- Experience
- 3+ years building models for business application; PhD, or Master's degree and 4+ years of CS, CE, ML or related field experience.
- Education
- PhD, or Master's degree and 4+ years of CS, CE, ML or related field experience. A PhD in machine learning, computer science, electrical engineering, or a related field is preferred.
About the job
Develop and optimize foundation models and systems for efficient, reliable long-horizon agentic workloads, including planning, tool use, and multi-step task execution.
What you'll do
- Design, train, fine-tune, evaluate, and improve custom foundation models.
- Optimize training and inference; profile and address compute, memory, communication, and distributed-training bottlenecks.
- Develop evaluations for model reliability and long-horizon agent performance.
What they ask for
- Strong background in machine learning, deep learning, transformers, and foundation models; experience with PyTorch, JAX, or similar frameworks.
- 3+ years of building models for business application experience.
- PhD, or Master's degree and 4+ years of CS, CE, ML or related field experience.
- Experience with patents or publications at top-tier peer-reviewed conferences or journals; programming experience in Java, C++, Python or related language; experience in at least one listed area such as parallel and distributed computing or high-performance computing.
Nice to have
- Experience using Unix/Linux and in professional software development.
- Experience with AWS Trainium, Inferentia, Neuron SDK, TPUs, or other custom accelerators; PhD in a related field.
Pay
171,600.00 - 222,200.00 USD annually (base salary)
Tools
PyTorch, JAX, CUDA, Triton, AWS Trainium, Inferentia, Neuron SDK, TPUs
AI focus
Build foundation models for long-horizon agentic workloads, supporting planning, tool use, memory, self-correction, and multi-step execution; improve model efficiency, stability, and reliability.
Model-written summary · the original posting is authoritative