Applied Scientist: AI-Driven Ops & Forecasting

Amazon Science; Bellevue (WA); On-site; USD 136,000 - 184,000; Full time

Benefits offered by this job

Health insurance

RSUs

Paid time off

Job summary

Amazon Science is seeking an Applied Scientist to advance the operational excellence of predictive models within Amazon's North American Operations. You will own projects end-to-end, collaborating with stakeholders across Operations and Finance, and deploy production-grade models in cloud environments with CI/CD practices.

The role emphasizes building explainable AI systems and robust evaluation frameworks to detect quality regressions before customers experience them.

Qualifications

• Master's degree or higher in computer science, ML, engineering, or related field.
• Experience programming in Java, C++, Python or related language.
• Experience with SQL and an RDBMS (e.g., Oracle) or Data Warehouse.
• Experience in building machine learning models for business application.

Responsibilities

• Interfacing with stakeholders to understand business problems and translating these into high and low-level design documents.
• Architecting, building, and deploying production-grade models to cloud-native environments, incorporating CI/CD best practices.
• Designing AI systems to provide model explainability, as well as automate customer workflows.
• Developing evaluation frameworks for models that identify quality regressions before customers do.

Skills

Master's degree or higher

Java

Python

SQL

ML models

Education

Master's degree or higher in CS/ML/Engineering

Tools

Oracle

Data Warehouse

Job description

Amazon Science is seeking an Applied Scientist to advance the operational excellence of predictive models within Amazon's North American Operations. You will own projects end-to-end, collaborating with stakeholders across Operations and Finance, and deploy production-grade models in cloud environments with CI/CD practices.

The role emphasizes building explainable AI systems and robust evaluation frameworks to detect quality regressions before customers experience them.