Design agentic AI systems for Alexa Daily Essentials, applying reinforcement learning and evaluation science to customer-facing features at scale. Work across prototyping and production deployment.
As an Applied Scientist II, you will create systems that reason, plan, and act over complex data. You will apply reinforcement learning, retrieval-augmented generation, and rigorous evaluation methods. Your work will prototype AI-powered features, validate architecture decisions, and influence product roadmaps through experiments and measured quality. You will collaborate directly with engineers and product leaders while owning science end-to-end, from concept through launch. This team is not training foundation models;
it is advancing how they are applied through agentic architectures and evaluation.

Key Responsibilities

Design agentic systems for reasoning, tool use, planning,



and orchestrationApply reinforcement learning to improve routing and task resolutionBuild evaluation frameworks, benchmarks, and regression detection pipelinesPrototype AI-powered features from concept through production launchDesign experiments that shape architecture and product roadmapsRequirements:
Three-plus years building machine learning models for business applicationsPhD, or master's degree and four-plus years of relevant experiencePatents or publications in top-tier peer-reviewed venuesProgramming experience in Java, C++, Python, or related languagesExperience with machine learning, reinforcement learning, or related applicationsBring applied science, reinforcement learning, and evaluation expertise to build agentic Alexa experiences serving millions of customers daily.

📌 Applied Scientist Agentic Ai Systems (Canada)
🏢 Amazon
📍 Canada