Applied Scientist, Leo Security

About the Role

Amazon Leo is seeking an Applied Scientist to join the founding cohort of the Engineering and R&D team within Leo Infrastructure and IP Security. This team is responsible for defending the manufacturing lines, launch sites, and global ground infrastructure of the Leo constellation from sophisticated threat actors. The role involves building statistical and behavioral models to differentiate threat actor behavior from operational noise, and developing privacy-preserving data representations. This is an R&D role with a production mandate.

Key Job Responsibilities

• Build behavioral and statistical models that baseline normal activity across heterogeneous security telemetry, including specialized industrial-control and factory-floor data sources, to enable low false-positive rate detections in new environments.
• Design privacy-preserving representations of sensitive security data and verify model/detection accuracy against them.
• Define methodologies and lead analysis for complex, loosely defined problems by gathering cross-domain data, selecting appropriate data science techniques, and justifying approaches with evidence.
• Develop metrics and evaluation frameworks to measure model and detection performance against threat actor behavior prior to production deployment.
• Contribute to the team's neurosymbolic reasoning platform by adapting state-of-the-art techniques and shipping production-quality components.
• Document work with rigor, akin to a peer-reviewed publication, and communicate findings to both scientific and security-operations audiences.

Requirements & Tech Stack

• Fluency in Python.
• Experience with statistical modeling and analysis of large, heterogeneous datasets, including anomaly detection, time-series analysis, or behavioral baselining.
• Experience with popular deep learning frameworks (e.g., PyTorch, TensorFlow).
• Experience with building models for business application.
• A PhD or a Master's degree and 4+ years of experience in Computer Science, Computer Engineering, Machine Learning, or a related field.
• Export Control Requirement: Candidates must be a U.S. citizen.

Qualifications

• Experience with a systems language such as Rust or C++ is a plus.
• Experience in one or more of the following domains: access-control systems and methodology, network security, application- and system-development security, security architecture and models, cryptography, and operations security.
• Experience in professional software and systems development.
• Publications at top-tier peer-reviewed machine learning, data mining, or security venues (e.g., NeurIPS, ICML, ICLR, KDD, CCS, USENIX Security).
• Experience with privacy-preserving machine learning, including synthetic data generation, anonymization, or differential privacy, and validating the fidelity of transformed data.
• Experience designing evaluation frameworks and metrics for models where ground truth requires human judgment.
• Experience modeling operational or sensor telemetry, including industrial control systems (OT/SCADA), IoT, or physical-world event data.