Amazon
Sr. Applied Scientist - Agentic AI, Amazon Connect
$167K–$226K
Seattle, WA · Senior · Full-time · Posted Oct 16, 2026
- Experience
- 5+ years building machine-learning models for business applications; PhD, or Master's degree and 6+ years of applied research experience.
- Education
- PhD, or Master's degree and 6+ years of applied research experience.
About the job
Researches, develops, evaluates, and deploys agentic AI science solutions for Amazon Connect, using measurement and experimentation to connect model quality with customer outcomes.
What you'll do
- Conduct research and develop machine-learning science solutions from conception through deployment.
- Define offline evaluation harnesses and online experiments, and establish metrics tied to customer outcomes.
- Prototype, iterate on, and evaluate robust agentic systems with modern tooling and AWS services.
What they ask for
- 5+ years of experience building machine-learning models for business applications.
- PhD, or Master's degree and 6+ years of applied research experience.
- Programming experience in Java, C++, Python, or a related language.
- Demonstrated experience with generative AI and LLMs, such as fine-tuning, prompting, retrieval, or post-training.
Nice to have
- Experience designing evaluation frameworks for open-ended or subjective tasks, including synthetic data generation.
- Experience with agentic and multi-turn systems, including RL-based post-training, environment simulation, or agent harness evaluation.
Pay
USD 167,100.00 - 226,100.00 annually (base salary)
Tools
PyTorch, vLLM, agent frameworks, MCP, Amazon Bedrock, SageMaker, Java, C++
AI focus
Research and develop agentic AI systems using generative AI and LLMs; build evaluation methods and metrics, and prototype, test, and deploy robust systems.
Model-written summary · the original posting is authoritative