Senior Applied Scientist - AWS Data Catalog Permissions

Role Overview

This role provides an opportunity to work at the forefront of data security and governance affecting AWS customers at scale. As a Senior Applied Scientist III specializing in AWS Catalog permissions, you will leverage your expertise in formal methods, programming languages, and automated reasoning to develop the permissions model that precisely governs data access. An ideal candidate brings experience with machine learning, including LLMs and multi-agent workflows, enhancing determinism in language model outputs, in a collaborative and cognitive-safe work setting.

Key Responsibilities

• Design and define a precise, compact, and expressive language and formal semantics allowing customers to specify permissions at granular resource levels such as catalog, database, table, column, and row.
• Ensure the permissions model remains correct, explainable in authorization denials, and exhaustively testable while balancing potential ambiguities.
• Document design decisions clearly and quickly align stakeholders around the approach.
• Lead development of scientifically challenging components that provide substantial customer benefits and are stable post-deployment.
• Address complex problems such as verifying policy changes do not widen access, equating policies across models, aligning fast evaluation paths with reference semantics, and maintaining permission intent through data and schema evolution.
• Maintain deep system knowledge and proactively drive cross-team improvements in efficiency and consistency.
• Collaborate with Applied Scientists, Engineers, and Product Managers across identity, storage, analytics, and machine learning to shape science and business strategies, and contribute to roadmap planning.
• Facilitate consensus-building among differing viewpoints with thoughtful feedback.
• Tackle bottleneck challenges using advanced techniques and establish a reputation as an innovative problem solver.
• Contribute to AWS's academic presence through publications, talks, and advancements in data security, governance, programming languages, and automated reasoning.

About The Team

The AWS Catalog permissions team is accountable for multiple critical areas: the design of permission models and formal semantics for granular resource authorization, providing the authorization decision evaluation path that adheres to strict latency requirements, detecting and resolving divergences between stated intentions and enforced permissions as resources evolve, and unifying multiple existing permission models into one while guaranteeing access controls are retained. Collaboration with AWS teams specializing in identity, storage, and analytics is essential.

Qualifications

• Minimum of 6 years experience developing machine learning models in business applications.
• PhD or Master's degree paired with at least 6 years of applied research experience.
• Proficiency in programming languages such as Java, C++, Python, or related languages.
• Experience with neural deep learning methods and machine learning technologies.

Preferred Experience

• Familiarity with modeling and data science tools including R, scikit-learn, Spark MLLib, MxNet, TensorFlow, NumPy, and SciPy.
• Experience working with large-scale distributed systems such as Hadoop and Spark.

Additional Information

Amazon is an equal opportunity employer committed to diversity and inclusivity. Accommodations are available for applicants with disabilities throughout the hiring process. Compensation includes a base salary with potential sign-on bonuses and restricted stock units, alongside comprehensive benefits such as medical, dental, vision, life insurance, retirement plans, and paid time off. Salary range for this position in Vancouver, Canada is CAD 195,900 to CAD 327,200 annually.