Find your next role

Description

The AI Center of Excellence (AICE) builds AI primitives that power system-intrinsic intelligence and Trusted Intelligent Knowledge Infrastructure for both AI-as-a-Consumer and human users. We develop AI capabilities independently and in partnership with product-owning teams to evolve Amazon's AI-driven operations and productivity. Our impact is enterprise-broad and global.

We are looking for an Applied Scientist who can identify high-impact problems, design rigorous experiments, and translate research into hardened, reusable primitives ready for broad consumption across the enterprise.

Key job responsibilities
* Research, design, and develop AI/ML primitives - defining problem spaces, formulating approaches, running experiments, and validating outcomes with scientific rigor
* Translate research findings into production-ready primitives, working closely with machine learning engineers to harden and scale solutions
* Drive the scientific direction of AICE primitives, identifying opportunities where novel approaches can unlock step-function improvements
* Design and execute experiments to validate primitive performance, robustness, and generalizability before broad consumption by partner teams
* Collaborate with product-owning teams to understand their problem spaces and ensure AICE primitives deliver measurable value when integrated
* Contribute to AICE's mission of driving AI transformation through best practices in applied research, evaluation methodology, and responsible AI within your area of expertise
* Publish findings, share learnings, and raise the scientific bar across the team

A day in the life
Every day brings new challenges. You might be deep in experimentation - designing evaluations to stress-test a new primitive's robustness. The next day, you could be analyzing results and iterating on model architectures. Later in the week, you may collaborate with partners to understand their domain and shape a primitive that fits their product needs. Other days, you'll work alongside our MLEs to harden a validated approach for broader consumption.

No two weeks look the same. You'll move fluidly between research and applied problem-solving - always with the goal of building AI primitives that are scientifically sound, trusted, and impactful at enterprise scale.

About the team
AICE is a team of scientists and machine learning engineers building the AI primitives that power Amazon's intelligent systems. We develop and harden reusable capabilities for broad consumption by product teams across the enterprise - and partner with those teams to integrate them.