About the job
Join us in building Reflex, the agentic incident-response platform for Amazon's fulfillment network. You'll design and ship production software and AI agents on Amazon Bedrock AgentCore that triage high-severity incidents, generate real-time call intelligence, draft stakeholder communications, and produce structured post-incident records, shifting incident response from a manual, pull-based model to an intelligent, push-based one.
Amazon's network of fulfillment centers is the infrastructure that Amazon Robotics runs on. When it degrades, robots stop and packages stop moving. The team manages thousands of high-severity incidents every year, with Incident Managers assembling context across many systems under time pressure before resolution work can even begin. The software you build puts that context in front of responders in seconds and removes the repetitive work that extends incidents today. You'll design the data, evaluation, and feedback mechanisms that make the system measurably better with every incident it touches.
This role sits on a software team within Operations Infrastructure Services (OIS), part of Amazon Robotics. You'll stay close to live operations through incident reviews, workflow observation, and call shadowing, and turn what you learn into durable software. Reflex is the immediate focus; as it matures, the same foundations are expected to extend toward shared incident context across organizations, coordinated agent workflows, and carefully guarded automation of recovery validation and repeatable response actions, with every step gated by measurable confidence.
Responsibilities
- Design, build, test, deploy, and operate production services and AI agents on AWS (Amazon Bedrock AgentCore, serverless compute, event-driven pipelines) that automate incident triage, call intelligence, communications, and post-incident documentation and reporting
- Own features end-to-end: from discovery with Incident Managers and resolver teams, through design, implementation, evaluation, deployment, and production operation
- Build the foundations that gate agent autonomy: LLM output evaluation, observability and alerting for agents in production, and identity and access controls aligned with Amazon standards
- Design the feedback loops through which agents learn: capturing human reviews, corrections, approvals, and incident outcomes as evaluation signal, and turning resolved incidents into structured history that improves recommendations over time
- Integrate with the incident lifecycle (ticketing, chat, telemetry, detection feeds, and live call transcription) and model consistent incident state across those systems
- Raise the bar on software quality, security, testing, and operational excellence for AI systems acting inside production incident workflows
Qualifications
Minimum
- 3+ years of non-internship professional software development experience
- 2+ years of non-internship design or architecture (design patterns, reliability and scaling) of new and existing systems experience
- 1+ years of software development engineer or related occupational experience
- 1+ years of designing and developing large-scale, multi-tiered, multi-threaded, embedded or distributed software applications, tools, systems, and services using: C#, C++, Java, or Perl experience
- 1+ years of Object Oriented Design experience
- Bachelor's degree or foreign equivalent in Computer Science, Engineering, Mathematics, or a related field
- Experience programming with at least one software programming language
Preferred
- 3+ years of full software development life cycle, including coding standards, code reviews, source control management, build processes, testing, and operations experience
- Knowledge of Machine Learning and LLM fundamentals, including transformer architecture, training/inference lifecycles, and optimization techniques