We are responsible for billions of dollars of payments each month to millions of Amazon vendors across the globe. We delight our vendors by paying them when they want and how they want, accurately and on time. We build software applications that facilitate global financial transactions worth billions (USD) annually, rivaling the complexity and scope of dedicated payments fintech companies. Our systems decide how and when a payment happens, carry it out across the world's payment networks, and account for it — all under real regulatory, tax, and audit obligations, in every geography Amazon operates in. You will have the opportunity to work in a high-paced environment leveraging the AI-Driven Development Lifecycle (AIDLC) within our core principles of engineering, to deliver faster and ship features incrementally. Ownership is the center of this role, and ownership here means the whole lifecycle rather than just the code. You will take features from working backwards from the customer through design, implementation, testing, deployment, and documentation — and you stay accountable for how they behave in production long after they ship. Own features and components of our payment systems end-to-end. The business problem and the technology direction will be understood; how to actually build it will not be, and that part is yours. Design the solution and make the tradeoffs. You will code independently, and you will have senior and principal engineers to pull in when a tradeoff is genuinely hard or when a design needs to line up with where the architecture is heading. Work directly with product managers and business stakeholders — finance, tax, and accounts payable operations — to understand what they need and why, so that you build the right thing. You will hear the requirement first-hand rather than through a written hand-off, and you will be expected to push back when the ask and the outcome do not line up. Ship incrementally. Large payment changes land safely through staged rollout and measurement, not through a single release. Reduce operational load. Our best work turns a recurring manual intervention into a code fix or a self-healing mechanism, not another runbook step. Leave the system better explained than you found it. What you build should be maintainable and extensible by an engineer who has never seen it, which means clear documentation and design choices that hold up without you in the room. Raise the bar around you: code review as a teaching mechanism, mentoring engineers earlier in their careers, and getting new teammates productive on how our software is built, how it operates, and how it fits into the bigger picture. Building with agentic AI We are not experimenting with AI on the side — it is how we work. You will use AI agents across the development lifecycle: exploring unfamiliar systems, drafting and reviewing designs, generating and testing code, triaging production issues, and automating the operational work that used to consume a week of on-call. We expect engineers to be strong directors of these agents and rigorous verifiers of their output. The judgment stays with you: you are accountable for what ships, whether you typed it or an agent did. Agentic capabilities are increasingly part of the systems themselves — automating decisions and investigations that previously required a human in the loop — and you will have room to build them. Why this team Payments is one of the few places in software where the business logic is genuinely hard, the scale is enormous, and a mistake has an immediate real-world consequence — a vendor who does not get paid when they expected. If you want real ownership, global scope, and the freedom to build with the newest tools available rather than waiting for permission to try them, this is a good place to build a career.