Applied Scientist – AGI & LLM Training Environments

Amazon Science; Sunnyvale (CA); On-site; USD 172,000 - 222,000; Full time

Job summary

Amazon Science is seeking an Applied Scientist with a strong deep learning background to help build training environments for LLMs. You will design RL gyms, define tasks and rewards, and scale infrastructure to improve model capabilities.

You’ll work closely with model training teams to ensure environments drive real value in generative AI applications. The Foundational AI (FAI) team is pushing GenAI boundaries to deliver the best customer experiences, with a focus on robust, scalable ML systems

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 programming in Java, C++, Python or related language.
• Experience in algorithms, data structures, parsing, numerical optimization, data mining, parallel and distributed computing, HPC.

Responsibilities

• Design, build, and evaluate reinforcement learning environments (RL gyms) for LLM training.
• Define tasks and reward signals, and build infrastructure to run environments at scale.
• Measure environment transfer to model capability and verify quality and correctness.
• Collaborate with model training teams to maximize environment value and impact.

Skills

Java

C++

Python

Algorithms and data structures

Numerical optimization

Distributed computing

Education

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

Tools

Unix/Linux

Job description

Amazon Science is seeking an Applied Scientist with a strong deep learning background to help build training environments for LLMs. You will design RL gyms, define tasks and rewards, and scale infrastructure to improve model capabilities.

You’ll work closely with model training teams to ensure environments drive real value in generative AI applications. The Foundational AI (FAI) team is pushing GenAI boundaries to deliver the best customer experiences, with a focus on robust, scalable ML systems