Software Development Engineer II, Selection Monitoring

Help Amazon understand every brand in the world and make sure customers can find the brands they love in our store. As a Software Development Engineer on the Universal Brand Catalog (UBC) team, you'll design and own large-scale data systems, AI/ML-powered services, and self-serve tools that show brand recruitment teams which brands are missing from Amazon and where to act. This is a new space with key architecture decisions still on the table, so you'll have real influence over the technical direction from day one. You'll work across the full software development lifecycle, collaborate closely with applied scientists and product managers, and see your work translate directly into wider selection for Amazon customers. Key job responsibilities - Design, build, and operate scalable, highly available services and large-scale data pipelines (Spark on Amazon EMR) that process brand, product, and marketplace data across Amazon's global catalog. - Own features end to end — requirements, design documents, implementation, testing, deployment, and operations — and lead design reviews where you make sound tradeoffs between accuracy, latency, cost, and scalability. - Partner with applied scientists to put ML models into production, measure data quality through precision and recall metrics, and iterate on brand-matching and gap-detection signals. - Turn ambiguous business problems into clear technical plans, break large projects into shippable deliverables, and build self-serve tools that let business teams slice brand and selection data by market, category, and competitive trends without engineering support. - Mentor junior engineers and interns through code reviews, design guidance, and pairing, while driving operational improvements in monitoring, alarming, and on-call health. A day in the life You might start the day at stand-up, then review a design document for a new pipeline that scores brands across marketplaces. Later you could meet with a product manager and a scientist to figure out how to measure whether a brand gap signal can be trusted, then turn that into a data model and plan. In the afternoon you'll write code, review a teammate's pull request, and help a junior engineer debug a Spark job. Your internal stakeholders include category and vendor management teams who use your data to recruit brands, partner catalog teams, and science teams. About the team The UBC team builds the foundation for how Amazon understands brands: which brands exist, how they relate to each other, where they're sold, and where they're missing from our store. We own Brand Gaps (finding brands and selection absent from Amazon) and Brand Cohort Management (grouping and prioritizing brands so teams can focus on what matters most to customers). Engineers and applied scientists work side by side on problems like signal precision, cohort ranking, and entity matching. The space is new and growing, so you'll have real ownership, see the business impact of your work directly, and grow through mentorship, design reviews, and a sustainable on-call culture.