Amazon Advertising is a rapidly expanding and highly profitable sector within Amazon, dedicated to creating advertising solutions that boost product discovery and sales. Our systems process billions of ad impressions daily, generating substantial data across the entire advertising lifecycle.

We are developing an AI-powered conversational assistant, SpektrBot, designed to enable data engineers, analysts, and business stakeholders to interact with our extensive advertising data lake using natural language. By leveraging cutting-edge generative AI, SpektrBot simplifies insight generation, transforming natural language queries into precise data retrievals and bridging the gap between questions and answers.

This role offers a unique chance to build a 0-to-1 product at Amazon's scale, situated at the convergence of generative AI, data engineering, and conversational interfaces. If you are passionate about constructing production-grade AI systems, comfortable with ambiguity, and eager to see your work utilized daily across Amazon Advertising, we encourage you to apply.

- Architect and implement scalable backend services to power SpektrBot's conversational AI platform, including query orchestration, context management, session handling, and response generation. - Develop Retrieval-Augmented Generation (RAG) architectures, optimizing retrievers, vector stores, embedding pipelines, and metadata indexing for accurate data retrieval from our advertising data lake. - Integrate and operationalize Large Language Models (LLMs), building infrastructure for prompt management, model routing, multi-agent orchestration, and chain-of-thought workflows to support natural, multi-turn conversations. - Create automated evaluation and feedback systems to continuously monitor response accuracy, detect regressions, and implement corrections for ongoing system improvement. - Engineer SQL generation and validation pipelines, managing the complete flow from natural language intent to SQL output, incorporating guardrails, validation, and result verification for trusted results. - Develop LLM-powered tools for automatic metadata enrichment, classification, and maintenance, enhancing the metadata catalog for precise data retrieval. - Partner with applied scientists to productionize NLP/NLU models and integrate them seamlessly into the bot's architecture. - Take full ownership of systems from design through deployment, monitoring, alarming, and operational excellence within a production environment serving thousands of internal customers.

- Possess over 3 years of professional software development experience, excluding internships. - Have more than 2 years of experience in designing or architecting new and existing systems, focusing on reliability and scalability using design patterns. - Offer at least 1 year of experience in designing and developing large-scale, multi-tiered, multi-threaded, embedded, or distributed software applications, tools, systems, and services utilizing languages such as C#, C++, Java, or Perl. - Demonstrate at least 1 year of experience with Object-Oriented Design principles. - Proficient in programming with at least one software programming language. - A Bachelor's degree in Computer Science or a related field is preferred.