Software Development Engineer, Measurement, AdTech, and Data Science

Build the attribution systems that connect ad interactions with real customer outcomes — processing more than 50 billion events per day to produce the measurements that power experiences across all of Amazon Ads and help millions of advertisers understand the true value of their campaigns. Attribution is at the heart of advertising measurement, and your work will directly influence how advertising budgets are optimized at global scale. Building accurate attribution at Amazon scale is one of the most challenging engineering problems in advertising. Our systems connect signals across advertising products, customer journeys, devices, channels, and conversion sources — solving problems involving massive-scale data processing, distributed systems, identity and signal fragmentation, evolving privacy requirements, and complex attribution methodologies. The measurements we produce provide foundational signals for reporting, campaign optimization, machine learning models, analytics applications, and emerging AI-powered advertising experiences. What makes this role stand out: • 50B+ events per day — petabyte-scale data pipelines using Spark, EMR, Apache Iceberg, DynamoDB, and other AWS services • Engineering meets science — you'll work directly with applied scientists and economists to translate new measurement methodologies into scalable production systems • End-to-end ownership — from understanding an ambiguous measurement problem through design, implementation, launch, and operations • Direct business impact — the data you produce helps millions of advertisers make better investment decisions and powers optimization across Amazon's full advertising suite • Build new, not just maintain — opportunities to design new architectures that improve scalability, data freshness, correctness, reliability, and cost efficiency Key job responsibilities - Design, develop, and deliver software that produces measurement data consumed across Amazon's full advertising suite — spanning display, search, native, and video - Build and maintain petabyte-scale data pipelines and services using technologies like Spark, EMR, Apache Iceberg, and Java, inventing new big data paradigms to keep pace with rapid growth - Navigate complex attribution challenges with creative problem-solving — adapting to evolving technical requirements, privacy landscapes, and shifting priorities - Collaborate with applied scientists, economists, product managers, and engineers across teams to translate measurement methodologies into scalable production systems - Apply a variety of architectural approaches and design patterns to deliver maintainable, scalable, and well-tested solutions in a high-throughput environment - Drive engineering best practices including code reviews, testing strategies, operational excellence (alarms, runbooks, monitoring), and continuous improvement - Contribute to technical design discussions, identify trade-offs, and propose solutions that balance performance, cost, and long-term maintainability About the team The Measurement, AdTech, and Data Science (MADS) team defines and produces the metrics advertisers rely on to analyze the performance of their ad investments. Our charter is focused on the systems responsible for computing and distributing estimated conversions at massive scale. We work across a rich technology stack — combining AWS services like EMR, Kinesis, and AWS Batch with open-source technologies such as Spark and Presto — to deliver fast, reliable measurement at the speed advertisers demand.