About the job
Amazon is looking for a passionate and inventive scientist to advance the science in foundational models and Agentic AI. Specifically, as part of our science team in Amazon AWS Agentic AI, you will lead the research and development of techniques for efficient and effective Agent optimization/learning, through various techniques ranging from model fine-tuning (e.g. RL) to context optimization, as a foundational layer for AWS customers to build reliable and performant Agents.
Responsibilities
Develop short-term and long-term science roadmap for research in the broad area of Agent optimization/learning, which could range from RL reward shaping, training efficiency optimization, to automatic prompt optimization and techniques to learn from agent memory.
Develop concrete science plan, implement and validate research idea before moving it to production
Collaborate across product and engineering teams to transfer science innovations into AWS customer facing product.
Qualifications
Minimum
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
Preferred
Experience with techniques like RLVR, reward modeling, inference optimization, agent task/environment synthesis.