Principal Applied Scientist, Conversational Assistant Modeling & Learning
Amazon · Bellevue, WA · Applied Science
About this role
Amazon is hiring a principal-level Applied Scientist in the machine learning function based in Bellevue, WA. The posting calls out experience with Python, Java, LLMs, Reinforcement Learning and roughly 5+ years of relevant work. Listed education preference: a Ph.D. or equivalent. Compensation is listed at $198,900–$269,000 per year.
- Role
- Applied Scientist
- Function
- machine learning
- Level
- principal
- Track
- Tech leadership
- Employment
- Full-time
- Location
- Bellevue, WA
- Experience
- 5+ years
- Education
- Ph.D. preferred
- Department
- Applied Science
- Posted
- Mar 19, 2026
More roles at Amazon
Job description
from Amazon careersAlexa AI is looking for a Principal Applied Scientist to lead the science behind Alexa+, Amazon's LLM-powered conversational assistant. You will own the technical direction for key initiatives spanning large language model fine-tuning, alignment, agentic reasoning, and evaluation — directly shaping the experience for hundreds of millions of customers worldwide. As a Principal Scientist, you are a hands-on technical leader. You define research directions, design and run rigorous experiments, and ensure that research translates into production systems at scale. You decompose ambiguous, hard problems into clear solutions. Your code, models, and documents are exemplary and frequently referenced across the organization. You amplify your impact beyond your own work. You lead scientific reviews, scrutinize experimental design and modeling assumptions, and align teams toward coherent strategies. You mentor senior scientists, contribute significantly to hiring, and keep the broader scientific community current on state-of-the-art techniques. You bring business and industry context to technical decisions and can credibly present to executive leadership. Key job responsibilities Define and drive the science roadmap for conversational AI capabilities powered by large language models Design, implement, and evaluate novel approaches to LLM fine-tuning, alignment (RLHF, DPO), and distillation for production deployment Architect agentic systems — multi-step reasoning, tool…