Software Development Engineer, Amazon Customer Service

When a customer's Kindle won't sync, their Prime Video subscription charged them twice, or an audiobook never downloaded, they come to us. Nine times out of ten they would rather fix it themselves than wait for an agent — and our job is to make that possible. CX-STAR (Subscriptions, Troubleshooting, and Accounts Resolutions) builds the self-service support experience for Amazon's devices and digital products: Kindle, Fire TV, Echo, Alexa, Prime Video, Amazon Music, and Kindle Unlimited. We are hiring software development engineers to build it. You will work across the stack — React front ends customers touch, Java services behind them, AWS infrastructure underneath — and on the large language model work now moving into production in our support flows. Every feature you ship runs as a live experiment, so you will know within weeks whether it actually helped, and you will help decide what happens next. Key job responsibilities - Own features end to end on Amazon's customer service pages for devices and digital products — design, implementation, testing, deployment, and operations - Design solutions for problems where the goal is clear but the approach is not, and bring clarity to work that carries real technical risk - Build across a React component library, Java services, and AWS infrastructure — refactoring and simplifying as readily as you add - Own the A/B experiment that launches your feature: read the results, weigh the customer-impact and performance data, and recommend whether it ships to everyone, holds, or rolls back - Launch features into new countries, accounting for translation, regional data differences, and per-country release requirements - Carry on-call for what you own, drive failures to a permanent root-cause fix, and make the case for operational work against new features - Review your teammates' code — including that of more senior engineers — mentor engineers newer to the team, and document your software well enough that someone who has never seen it can extend it A day in the life Mornings start with the sprint board and, on your on-call weeks, a look at pipeline health. Today it is a cluster of bugs from testers in France and Germany — all tracing back to one assumption that only held in the US. You find the shared cause and fix it once instead of twenty times. Midday brings a design call. Your feature needs a device action owned by another team, and there are two paths to it. You talk to them, learn what each really costs, and pick the one that ships. Then standup, a code review for a teammate, and your own change squashed into a single clean commit. In the afternoon your experiment from three weeks ago has enough data. Contacts are down in one marketplace and flat in another. You write up what you see and take a recommendation to your product manager: dial it to 100%, or hold two more weeks. That call is yours to make and defend. The loop is always the same, and it is the best part of the job: build it, measure it, decide what the measurement means.

Our team builds self-service support for Amazon's devices and digital products — Kindle, Fire TV, Echo, Alexa, Prime Video, Amazon Music, and Kindle Unlimited — as part of the CX-STAR organization. We are a small team across Austin and Seattle, spanning new-graduate hires through senior engineers with ten-plus years at Amazon. That range is deliberate: you will review a senior engineer's code and have your own reviewed by someone a level-up, in the same week. We run two-week sprints, share an on-call rotation, and ship to customers in ten countries. Product managers, quality engineers, and partner teams are in the room for design decisions, not handed a result afterward. And because every feature launches as a measured experiment, disagreements here get settled with data — a habit worth building early.