Applied Scientist, Ads Measurement Science

This role develops and productionizes end-to-end ads measurement science solutions for Amazon Ads. You will design and analyze models that provide measurement insights, implement critical-path code, and partner with engineers to bring solutions into production.

• Leverage deep expertise to invent solutions for ambiguous ads measurement problems
• Disambiguate problems to propose evaluation frameworks and success criteria
• Implement critical-path code and partner with engineers to productionize solutions
• Write high quality technical documents and communicate with technical and non-technical audiences
• Mentor junior scientists and contribute ideas to the team's direction

You will work on measurement solutions using scientific methods such as Generative AI, classical ML, Causal Inference, Natural Language Processing, and Computer Vision. The position emphasizes model evaluation frameworks, scalable implementation via critical-path code, and strong foundations in algorithms, parsing, numerical optimization, data mining, and parallel/distributed high-performance computing.

The ideal candidate is an applied scientist with 3+ years building models for business applications and strong programming experience in Java, C++, and/or Python. They have hands-on experience with at least one of Generative AI, classical ML, Causal Inference, Natural Language Processing, or Computer Vision, and can take measurement models from design through production with clear evaluation frameworks.