Lead Neurosymbolic Security Scientist

Amazon; Arlington (VA); On-site; USD 167,000 - 226,000; Full time

Job summary

Amazon Leo is building a foundational applied science team to defend its Low Earth Orbit constellation from threats. You will develop neurosymbolic reasoning platforms and models to detect sophisticated threat actors, bridging research with production deployment.

You will own the end-to-end science lifecycle, collaborate with security engineers, and shape the scientific agenda from day one. This is a high-impact R&D role with production mandates.

Qualifications

• 3+ years of building machine learning models for business application.
• PhD or Master’s degree with 6+ years of applied research experience.
• Proficient in Java, C++, Python or related language.
• Experience with neural deep learning methods and ML.
• Experience with graph algorithms, knowledge graphs, NLP, embedding retrieval.
• Experience designing evaluation frameworks with human judgment and A/B testing.

Responsibilities

• Design and implement scalable neurosymbolic systems with graph-based knowledge and LLM agents.
• Design and run reinforcement learning and fine-tuning pipelines for security reasoning.
• Build models to detect threat actor behavior and evaluate performance before production.
• Design multi-agent systems for triage, enrichment, and containment of security events.
• Own end-to-end science lifecycle from research to production deployment.
• Publish, patent, or open-source contributions to advance the field.

Skills

ML model development

PhD or Master with 6+ years

Java/C++/Python

Deep learning

Graph algorithms / knowledge graphs

NLP & embedding retrieval

Evaluation frameworks for AI systems

Education

PhD in a related field

Master's with 6+ years of applied research

Tools

TensorFlow

PyTorch

scikit-learn

Spark MLLib

MxNet

NumPy / SciPy

Job description

Amazon Leo is building a foundational applied science team to defend its Low Earth Orbit constellation from threats. You will develop neurosymbolic reasoning platforms and models to detect sophisticated threat actors, bridging research with production deployment.

You will own the end-to-end science lifecycle, collaborate with security engineers, and shape the scientific agenda from day one. This is a high-impact R&D role with production mandates.