Applied Scientist, Internal Audit
Are you passionate about data, enjoy solving complex analytical problems, leveraging industry leading agentic AI technologies to derive insight at scale - all in a challenging, fast-paced environment? We are seeking an Applied Scientist to accelerate the growth of Amazon Internal Audit’s Data Science & Risk Intelligence initiatives. The team builds ML and AI solutions that expand self-service data utilization by audit teams, utilizing the right methods to derive deeper patterns, and surface insights to gain holistic perspectives while amplifying potential risk mitigation.
Key job responsibilities
- Work with audit teams, product managers, engineers, and more senior scientists to deliver machine learning and generative AI products that carry real degrees of ambiguity, scale, and complexity.
- Design, build, and evaluate agentic AI systems — multi-agent workflows, retrieval-augmented generation, and tool-using agents — that automate and augment audit work, applying rigorous LLM-as-judge and human-aligned evaluation to measure and improve output quality.
- Apply statistical analysis and classical machine learning using SQL and scripting languages like Python/R over large datasets to develop insights and recommendations that strengthen internal audit.
- Architect secure, scalable solutions on AWS machine learning and generative AI services (e.g., Bedrock, AgentCore, SageMaker), owning the full lifecycle from prototype through production deployment, monitoring, and iterative improvement, working closely with auditors to understand their business needs.
- Build and maintain the team's production and experimentation infrastructure, including deployment pipelines, observability and tracing, and evaluation harnesses.
- Advance applied research by exploring emerging techniques and sharing findings through internal and external publications, talks, and conferences.
A day in the life
As an Applied Scientist, you will help shape and execute a product roadmap that connects risk to the business, building AI products — increasingly centered on large language models and agentic systems — that make audit work more effective and efficient. Your work spans the full arc of applied science: framing ambiguous problems, prototyping with the latest generative AI techniques, building rigorous evaluations, and deploying solutions in production. The ideal candidate pairs a strong foundation in data science and machine learning with a builder's instinct for production architecture, thrives on ambiguity, and stays close to a fast-moving research frontier.
About the team
Internal Audit’s mission is to help our businesses improve controllership, operational efficiency, and customer experience.
Key job responsibilities
- Work with audit teams, product managers, engineers, and more senior scientists to deliver machine learning and generative AI products that carry real degrees of ambiguity, scale, and complexity.
- Design, build, and evaluate agentic AI systems — multi-agent workflows, retrieval-augmented generation, and tool-using agents — that automate and augment audit work, applying rigorous LLM-as-judge and human-aligned evaluation to measure and improve output quality.
- Apply statistical analysis and classical machine learning using SQL and scripting languages like Python/R over large datasets to develop insights and recommendations that strengthen internal audit.
- Architect secure, scalable solutions on AWS machine learning and generative AI services (e.g., Bedrock, AgentCore, SageMaker), owning the full lifecycle from prototype through production deployment, monitoring, and iterative improvement, working closely with auditors to understand their business needs.
- Build and maintain the team's production and experimentation infrastructure, including deployment pipelines, observability and tracing, and evaluation harnesses.
- Advance applied research by exploring emerging techniques and sharing findings through internal and external publications, talks, and conferences.
A day in the life
As an Applied Scientist, you will help shape and execute a product roadmap that connects risk to the business, building AI products — increasingly centered on large language models and agentic systems — that make audit work more effective and efficient. Your work spans the full arc of applied science: framing ambiguous problems, prototyping with the latest generative AI techniques, building rigorous evaluations, and deploying solutions in production. The ideal candidate pairs a strong foundation in data science and machine learning with a builder's instinct for production architecture, thrives on ambiguity, and stays close to a fast-moving research frontier.
About the team
Internal Audit’s mission is to help our businesses improve controllership, operational efficiency, and customer experience.