Are you driven by innovation and complex problem-solving? At Infrastructure Reliability, we build scalable solutions that ensure the reliability of Amazon's critical systems. Our team develops and operates tools that detect and prevent outages to maintain high availability across Amazon's global infrastructure. Join us to architect solutions that directly impact millions of customers, with the resources and support to make meaningful contributions.
The team at Amazon is responsible for building intelligent and real-time insights into service-to-service communications, network traffic, and event correlation across hundreds of Amazon's critical fulfillment and robotics services. Our solutions support visibility into anomalous service behavior to prevent and quickly recover from incidents, ensuring high availability to keep the Customer Promise.
We are seeking a talented Senior Applied Scientist to invent the next generation of agentic observability solutions at Amazon scale. In this role, you will define, lead, and build the science behind intelligent systems that reason about complex network and infrastructure telemetry, autonomously detect anomalies, and drive automated remediation. You will own the scientific direction end to end, partnering closely with engineering, product, and Network Development Engineers to translate a long-term science vision into concrete research and delivery roadmaps.
Working backwards from the needs of our customers and operations teams, you will take the lead on ambiguous, high-impact problems where neither the problem nor the solution is well defined, and deliver production systems that improve infrastructure availability at scale. You will invent new methods, drive their adoption across multiple teams, and remain deeply hands-on with the hardest technical challenges.
Key job responsibilities
* Define and own the science vision for agentic observability, translating it into research and engineering roadmaps in partnership with product and engineering leaders.
* Build ML and agentic AI systems that autonomously detect, classify, and correlate infrastructure anomalies across network, compute, and service layers at Amazon scale.
* Design and develop models for event correlation, root cause analysis, and predictive failure detection using time-series analysis, graph-based methods, and deep learning.
* Own the agentic architecture for automated observability workflows, including planning, tool integration, long-horizon reasoning, and multi-agent orchestration.
* Define and curate the datasets and evaluation methodologies needed to train, benchmark, and continuously improve detection and classification systems.
* Stay deeply hands-on: write production-quality, critical-path code and build core components that take systems from prototype to launch.
Partner with Network Development Engineers and operations teams to ground science solutions in real-world infrastructure behavior and operational needs.
* Mentor scientists and engineers, raise the science bar, and represent the team in the internal and external scientific community through publications and presentations.
A day in the life
You will solve real-world problems by analyzing large-scale network telemetry and operational data, designing experiments and simulations, and developing ML models that detect and prevent infrastructure incidents. Your work requires close collaboration with engineers, Network Development Engineers, product managers, and operations leaders across the organization. You will prepare written and verbal presentations to share insights with audiences of varying technical sophistication, and you will iterate rapidly between research and production deployment.
