Senior Software Development Engineer , Stores Foundational AI - Rufus
Location
Seattle, WA
Job Type
Full-time
Category
other-general
Posted
June 03, 2026
Description
We are building foundational LLMs for Amazon Stores that fuse world knowledge with deep e-commerce understanding to power next-generation shopping experiences. These systems continuously learn from real-world customer interactions to become more helpful, personalized, and context-aware over time.
We are looking for builders who are passionate about large-scale systems, AI innovation, and customer impact. You will work at the intersection of distributed systems, machine learning infrastructure, and science to bring frontier research—especially in post-training and reinforcement learning—into production at Amazon scale.
Key job responsibilities
* Architect and build scalable ML infrastructure powering LLM training and post-training workflows, including supervised fine-tuning, reinforcement learning, and continuous learning from live traffic
* Transform real-world customer interactions into high-quality training signals, enabling continuous model imp...
We are building foundational LLMs for Amazon Stores that fuse world knowledge with deep e-commerce understanding to power next-generation shopping experiences. These systems continuously learn from real-world customer interactions to become more helpful, personalized, and context-aware over time.
We are looking for builders who are passionate about large-scale systems, AI innovation, and customer impact. You will work at the intersection of distributed systems, machine learning infrastructure, and science to bring frontier research—especially in post-training and reinforcement learning—into production at Amazon scale.
Key job responsibilities
* Architect and build scalable ML infrastructure powering LLM training and post-training workflows, including supervised fine-tuning, reinforcement learning, and continuous learning from live traffic
* Transform real-world customer interactions into high-quality training signals, enabling continuous model imp...