Amazon's finance teams rely on financial controls to keep spending accurate and financial reports audit-ready. You will build the systems that automate those controls, catch errors and waste before money is spent, and give finance analysts fast, trusted answers through AI tools built on cloud services and large-scale data. Key job responsibilities - Build services that check purchase requests and orders against finance rules in near real time, then correct or flag issues before approval. - Automate financial controls, such as reviews of fixed assets, leases, cost centers, open orders, so finance teams spend less time collecting data by hand and more time acting on exceptions. - Build data pipelines on AWS, including Lambda, StepFunctions, Glue, EMR, DDB, and S3, that feed C-Hub/dashboards finance teams use during month-end close and audit reviews. - Build AI agents on Amazon Bedrock AgentCore with the Strands SDK, including MCP tools, context and memory management, evaluation, harnesses, and guardrails that keep finance answers accurate. - Big data experience with Spark, StarRocks, or similar OLAP engines. - Build web applications and browser extensions in React and TypeScript that let finance users review exceptions, filter large datasets, and export results. - Turn defined business requirements into technical designs, review them with senior engineers, and take them from prototype to production. - Own your services in production. Write the tests, alarms, and automated rollbacks that let a change reach customers in hours, not days. Take part in on-call rotation. - Work directly with finance, accounting, product, and data science partners, and explain technical tradeoffs. A day in the life Most of your day goes to building. You might start by looking into an alarm from an overnight data job that feeds a finance control. Then you meet an accounting partner to design how a manual review could run automatically. After lunch, you write code and tests and ship the change through an automated pipeline the same day. Some weeks, you test a new AI model against an evaluation set to see if it gives better answers. Near month-end close, you answer finance users who want to know why a control flagged a transaction. About the team We are a team of software engineers who build the tools that keep Amazon's spending accurate and its financial reports audit-ready. Our customers are finance analysts, accountants, and employees across Amazon who buy goods and services. We own our products end to end, from data pipelines and services to the web apps and AI agents people use every day. We ship changes through automated pipelines and measure our work by the manual effort and errors we remove. We value clear writing, careful reviews, and engineers who ask why a process exists before they automate it.
