Applied Scientist Amazon
Seattle, WA

Job Description

We are the Catalog Fact Extraction team - a group of developers and scientists dedicated to ensuring the Amazon.com shopping experience gives our customers clean, consistent, and complete product information for every product in the world. We build thousands of machine learning models that extract and infer product information for every product, in every language, from every piece of text and image we sell. With hundreds of billions of facts to learn, we need applied scientists who are excited to use modern techniques to solve real-world problems.

Our customers need to compare specifications of televisions, find a dress with a specific color, or make sure their phone case fits a brand new device. These product facts make the Amazon customer experience possible, from search, discovery, and recommendations to exciting new shopping features. The models you build will leverage the latest big data, machine learning, cloud computing, and serverless computing technologies to find and extract these product facts from text and images.

You will help us build machine learning models that enable us to provide customers with a rich set of product facts. You'll modify existing techniques and watch your code fix billions of Amazon products, and then extend and invent new techniques using state-of-the-art research. You'll work side by side with engineers to make your models scalable, and have access to language experts to help you curate and evaluate training data across many languages. You'll see your techniques impact our customer experience and make an impact on people all over the world.

What else makes our team culture special? Our engineers and scientists work collaboratively on the same team, which means you're supported by a collaborative environment of peers all focused on solving the same problems for our customers. Our engineers love ML science, and will enthusiastically whiteboard, brainstorm, and prototype alongside you. You'll learn about how to apply science models at Amazon scale, and influence the engineering roadmap to accelerate science development.