Amazon
Applied Scientist, Agentic AI, AWS Governance
$143K–$193K
Seattle, WA +1 · Full-time · Posted Oct 26, 2026
- Location
- Seattle, WA
Arlington, VA - Experience
- 2+ years in each of the three basic-qualification areas; no alternatives stated.
- Education
- PhD in computer science, machine learning, engineering, or related fields is preferred.
About the job
Own and productionize machine-learning systems for AWS compliance verdicts, including NLI, evidence linking, and agent-context capabilities.
What you'll do
- Fine-tune, calibrate, deploy, and monitor the production NLI model and compliance framework pipelines.
- Develop evidence auto-linking and improve its ranking, embedding, reranking, and OCR-related performance.
- Establish evaluation, labeling, and retraining processes; research agent-context capabilities for retrieval, grounding, and hallucination measurement.
What they ask for
- 2+ years of building models for business application experience.
- 2+ years of programming in Java, C++, Python or related language experience.
- 2+ years of hands-on predictive modeling and large data analysis experience.
Nice to have
- PhD in computer science, machine learning, engineering, or related fields.
- Experience building agentic systems using LLMs and familiarity with agentic frameworks such as Strands Agent or LangGraph; knowledge of LLM techniques and post-training.
Pay
142,800.00 - 193,200.00 USD annually (base salary); 142,800.00 - 193,200.00 USD annually (base salary)
Tools
Java, C++, Python, Strands Agent, LangGraph, LLMs, NLP
AI focus
Build and maintain NLI models for compliance decisions; develop evidence-linking and agent-context systems involving retrieval, grounding, memory, and hallucination measurement.
Model-written summary · the original posting is authoritative