Software Development Engineer, Amazon Marketing Cloud Assistant

Application deadline: Sep 26, 2026

Build the AI-powered conversational agent that"s transforming how advertisers interact with their data-replacing SQL queries and complex UIs with natural language that simply works. You"ll work at the intersection of large language models, distributed systems, and petabyte-scale advertising data, shipping a high-visibility product at the forefront of generative AI in advertising.

We"re looking for a Software Development Engineer II who is excited to solve hard, novel problems: natural language understanding, data accuracy at scale, agent evaluation, and translating complex analytical workflows into intuitive AI-driven experiences.

In this role, you will:

• Design and build core backend services powering an LLM-based conversational assistant for Amazon Marketing Cloud
• Own end-to-end delivery of features-from system design through production deployment and operational excellence
• Build scalable data pipelines and infrastructure that ingest and process petabyte-scale advertising data from multiple sources
• Shape the future of how advertisers generate insights, build audiences, and optimize campaigns through AI

This is a fast-paced, ownership-driven environment where you"ll collaborate across science, product, and engineering to ship products that directly impact how millions of advertisers succeed on Amazon.

Who We"re Looking For

We seek someone with a strong coding foundation and experience building distributed systems at scale-someone energized by ambiguity and fast-moving problem spaces. Ideal candidates bring curiosity about LLM-based applications and a track record of delivering customer-facing products end to end. Experience with advertising technology, analytics platforms, or AWS services is a plus.

Key job responsibilities

AI Assistant Platform

• Design, develop, and maintain backend services powering the AMC conversational assistant
• Build and optimize LLM orchestration pipelines including prompt engineering, tool use, and agent workflows
• Develop APIs and integrations with AMC"s query engine, reporting services, and audience management systems
• Collaborate with applied scientists to integrate ML models and improve response quality
• Contribute to the evaluation framework-designing metrics, building ground truth datasets, and improving agent accuracy

Data Infrastructure

• Implement and maintain scalable data pipelines to ingest, process, and store petabyte-scale log-level advertising data from multiple sources (ad servers, exchanges, SSPs)
• Build ETL workflows that transform raw event-level data into queryable, analysis-ready formats following established architectural patterns
• Create and maintain APIs that enable analytics teams and data scientists to query log-level data efficiently
• Implement data quality checks, validation logic, and monitoring for data pipelines and services
• Optimize performance and cost efficiency of data processing jobs and queries

Engineering Excellence

• Own end-to-end delivery of features from design through production deployment
• Write clean, well-tested code and participate in code reviews to maintain high engineering standards
• Improve system reliability, latency, and scalability as the product scales to more advertisers
• Debug and resolve production issues, participating in on-call rotations as needed
• Document systems, processes, and runbooks to support operational excellence
• Collaborate with senior engineers to contribute to design discussions and architectural decisions

About the team

The Amazon Marketing Cloud (AMC) Assistant team is building an AI-powered conversational agent that helps advertisers interact with AMC through natural language. We are developing LLM-based experiences that simplify complex analytical workflows-enabling advertisers to generate insights, build audiences, and optimize campaigns without writing SQL or navigating complex UIs. Our work sits at the intersection of generative AI, advertising technology, and large-scale data systems.