Applied Scientist - Security AI & Neurosymbolic Systems

Amazon; Herndon (VA); Hybrid; USD 167,000 - 226,000; Full time

Benefits offered by this job

Health insurance

401(k) matching

Paid time off

Parental leave

Job summary

Amazon Leo seeks an Applied Scientist to be a founding member of the Engineering and R&D team within Leo Infrastructure and IP Security. You will build neurosymbolic reasoning platforms and models that detect sophisticated threat actors, with a production mandate guiding deployment.

You will frame scientific problems, prototype solutions, and ship capabilities that power the security platform at scale, collaborating across data, graphs, and LLM agents to deliver verifiable security outcomes.

Qualifications

• 3+ years of building machine learning models for business application
• Experience with graph algorithms, knowledge graphs, NLP, and embedding-based retrieval
• Experience designing evaluation frameworks for AI systems with human-ground truth and A/B testing

Responsibilities

• Design scalable neurosymbolic systems with graph-based knowledge and LLM agents
• Design and run RL and fine-tuning pipelines to optimize language models for security reasoning
• Build models to detect threat actor behavior and evaluate performance
• Develop multi-agent systems for triage, enrichment, and containment of security events
• Own the end-to-end science lifecycle from research to production
• Publishations, patents, or open-source contributions to advance the field

Skills

ML model development

Programming: Python/Java/C++

Neural networks / deep learning

Graph algorithms / knowledge graphs

NLP / embeddings

Evaluation frameworks for AI systems

Education

PhD

MS + 6+ years applied research

Tools

R

scikit-learn

Spark MLLib

MxNet

TensorFlow

NumPy

SciPy

Job description

Amazon Leo seeks an Applied Scientist to be a founding member of the Engineering and R&D team within Leo Infrastructure and IP Security. You will build neurosymbolic reasoning platforms and models that detect sophisticated threat actors, with a production mandate guiding deployment.

You will frame scientific problems, prototype solutions, and ship capabilities that power the security platform at scale, collaborating across data, graphs, and LLM agents to deliver verifiable security outcomes.