What if Amazon knew not just what you like, but what your partner is into, what to get your mom for her birthday, and what toy your dog will actually play with? We're working on that.
You'll work at the intersection of large language models and real-time personalization, solving problems that are genuinely open and at a scale very few teams get to work on. We have the foundation shipped; now we need to go deeper.
Key job responsibilities
As a Software Development Engineer, you will build systems at the intersection of AI, real-time data, and customer personalization at Amazon scale.
1. Design and build services that use large language models to extract structured knowledge from natural language and serve it in real time across Amazon's biggest customer surfaces.
2. Build reliable infrastructure around AI systems that are inherently unpredictable, handling hallucinations, latency spikes, and model drift so customers always get a great experience.
3. Develop pipelines that process millions of customer signals daily into durable, retrievable, personalized data stores.
4. Create monitoring and evaluation frameworks for AI outputs where traditional testing doesn't apply, including automated quality measurement and human-in-the-loop validation.
5. Collaborate with Applied Scientists to take models from research to production, owning the path from prototype to scaled deployment.
6. Build customer-facing feedback loops where real-world corrections compound system quality over time.
7. Ship fast and incrementally, owning the full lifecycle from infrastructure to alarms to on-call.
About the team
We are Amazon's Customer360 Personalization team. We build the system that understands every important person in a customer's life and personalizes Amazon for each of them individually. We identify what customers care about from signals across all of Amazon and curate it into personalized experiences for every person in their life. Our engineers are hyper-productive, leveraging AI daily to accelerate their own work while building systems that overcome AI's limitations at scale. The team values builders who take full ownership of what they ship, move fast, ask for help early, and care more about customer outcomes than technical elegance.
