PhD or equivalent Master’s Degree plus 4+ years of experience in CS, CE, ML, NLP, Statistics, Computer Vision, Applied Mathematics, or related field
2+ years of experience of building machine learning models for business application
Experience programming in Python, Spark, C/C++, or related language, as well as with R, Matlab or similar scripting language.
Communication and data presentation skills.
Ability to distill problem definitions, models, and constraints from informal business requirements, and to deal with ambiguity and competing objectives.
Experience in developing solutions for real-world applications.
Job summary
Are you interested in designing the artificially intelligent systems that guide Amazon Prime member experiences, world-wide? Would you like to create new scientific tools at the cutting edge of causal inference, machine learning, and distributed computing? If so, Amazon Prime Science is looking for you! We are seeking a talented applied scientist to join Prime’s core statistical science team.
Prime Science has two broad mandates. First, we design scalable systems for vending the right Prime member experience, at the right time, in real-time. Since Prime serves as Amazon’s premium membership, these systems interact with all Amazon customers around the world. Customers generate TB of data and billions of retail and grocery purchases. They also use our exclusive digital and entertainment content (e.g., Prime video, music, gaming, photos, and reading). Our second mandate is to design the Prime membership of the future. This involves using science to evaluate which new retail offerings or entertainment content to include as benefits to members. Modeling in this space powers company-wide profit allocation systems.
To contribute to these mandates, an applied scientist will partner with world-class i) scientists (applied, research, data scientists, and economists), ii) engineers (software, data, business intelligence), and iv) business professionals (product and program managers). They will design novel tools, leveraging and combining causal inference and machine learning. They will create scalable science from inception to production, using modern distributed computing software and infrastructure (e.g. EMR/Spark, Redshift, Sagemaker, DynamoDB, S3). Often, this involves creating new packages for customized scientific computing. They will present their work to other scientists, through internal and external conferences and journal submissions. They will patent their work. They will also present applied findings to business leaders.
Our team is unique in two primary ways. First, our work impacts systems across the entire company, given Prime’s broad scope. This makes our work integral to Amazon’s global business. Second, given our team’s mixed training in economics and machine learning, we specialize in combining tools from both paradigms. This includes the design of entirely new tools (e.g., estimators that combine inverse-reinforcement learning and structural econometrics), as well as novel applications of recently developed tools (e.g., generalized random forests or double de-biased machine learning). This is in addition to the application of more popular methods, like multi-armed bandits for content testing. An ideal candidate should have a passion for creating new science that spans interdisciplinary boundaries.
Primary responsibilities
Build and develop machine learning models and supporting infrastructure at TB scale, in coordination with software engineering and product teams.
Leverage state-of-the art tools for recommendation systems, including bandits, reinforcement learning, and inverse reinforcement learning.
Develop offline policy estimation tools and integrate with reporting systems.
Establish scalable, efficient, automated processes for large scale data analyses, model development, model validation and model implementation.
Work closely with business partners to understand problem spaces, identify opportunities and formulate solutions.
Design, develop and evaluate highly innovative models and statistical approaches to understand and predict customer behavior and to solve business problems.
PhD in Machine Learning, Computer Science, Statistics, Operations, Economics, Physics, Applied Mathematics or other highly quantitative field
Scientific publications in major journals and conferences
Experience handling terabyte size datasets
Knowledge of relational databases (SQL)
Experience working with distributed computing
Strong communication and data presentation skills
Amazon is committed to a diverse and inclusive workplace. Amazon is an equal opportunity employer and does not discriminate on the basis of race, national origin, gender, gender identity, sexual orientation, protected veteran status, disability, age, or other legally protected status. For individuals with disabilities who would like to request an accommodation, please visit https://www.amazon.jobs/en/disability/us.