RL Post-Training Scientist for LLMs

Amazon Inc.; Seattle (WA); Hybrid; USD 143,000 - 193,000; Full time

Job summary

Amazon Web Services, Inc. is seeking an Applied Scientist to advance reinforcement learning post-training for frontier large language models, focusing on instruction following, long-context reasoning, and tool use.

You will collaborate with world-class researchers, design experiments, publish results, and help scale research into practical AWS services, with access to vast data and compute resources. This role requires strong ML fundamentals, PhD or MS with research experience, and proven

Qualifications

• 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 in patents or publications at top-tier peer-reviewed conferences or journals.
• Experience programming in Java, C++, Python or related language.
• Experience in algorithms and data structures, parsing, numerical optimization, data mining, parallel and distributed computing, high-performance computing.

Responsibilities

• Work on Reinforcement Learning post-training for frontier LLMs to improve instruction following, reasoning over long context, and tool use.
• Collaborate with researchers and cross-functional teams to publish findings and contribute to AWS AI initiatives.

Skills

Java

C++

Python

Algorithms

Data mining

Parallel computing

Distributed computing

Research experience

Education

PhD or Master’s degree in CS/CE/ML or related field

Tools

Unix/Linux

Software development

Job description

Amazon Web Services, Inc. is seeking an Applied Scientist to advance reinforcement learning post-training for frontier large language models, focusing on instruction following, long-context reasoning, and tool use.

You will collaborate with world-class researchers, design experiments, publish results, and help scale research into practical AWS services, with access to vast data and compute resources. This role requires strong ML fundamentals, PhD or MS with research experience, and proven