Senior Software Development Engineer, Amazon Personalization - Customer 360
Summary
Architect and build a foundational AI system with persistent memory at scale, leading distributed retrieval and knowledge extraction for Amazon’s internal AI tools.
Key job responsibilities
As a Senior Software Development Engineer, you will own the technical architecture for core system components and lead the team through novel engineering challenges at the intersection of distributed systems, knowledge retrieval, and AI. You'll partner directly with Applied Scientists and leadership to translate research into production. Specific responsibilities include:
1. Own the end-to-end architecture for knowledge extraction, real-time retrieval, and cross-domain reconciliation systems.
2. Lead design and implementation of distributed systems handling millions of events per day with sub-second retrieval latency.
3. Make foundational choices about storage, indexing, and retrieval that will scale from day-one prototype to company-wide production.
4. Define technical standards, code review practices, and engineering culture for a team that's forming now.
5. Partner directly with Applied Scientists to translate research breakthroughs into production-grade systems.
6. Drive technical strategy with senior leadership visibility — your decisions shape the product roadmap.
7. Design systems that handle temporal decay, conflicting knowledge sources, and privacy-preserving retrieval at scale.
8. Mentor junior engineers and establish the patterns that the rest of the team will build on.
About the team
We're a new team within Amazon's Personalization organization, which built the recommendation systems that serve hundreds of millions of customers. We're applying that same depth of expertise to a fundamentally new domain — one where the "customer" is every builder at Amazon, and the "product" is an AI that compounds its value with every interaction.
The team is small, the problems are hard, and the autonomy is real. We value builders who invent patterns rather than follow playbooks, and who are energized by owning the full stack from research prototype to production system.