Software Development Engineer, Amazon Photos

Amazon Photos empowers customers to preserve, share, and relieve their memories across mobile applications, web, desktop, Amazon devices, and as part of Alexa+. We strive to provide customers with the best memories companion for them and their families. We're looking for a Software Development Engineer that wants to build and scale foundational services that power our computer vision and artificial intelligence capabilities. In this role, you'll architect and develop solutions that process millions of photos daily, enabling intelligent features that bring customer memories to life. You'll work at the intersection of large-scale distributed systems and AI technologies, creating services that form the backbone of Amazon Photos' next-generation features. Key job responsibilities - Design and implement scalable, fault-tolerant services that power computer vision and AI features across the Amazon Photos ecosystem - Build and optimize full-stack solutions that handle massive amounts of photo and video data while maintaining high performance and reliability - Architect and develop ML model serving infrastructure and pipelines for real-time computer vision applications - Drive technical decisions and architecture discussions for critical infrastructure components - Collaborate with ML scientists to productionize computer vision models and optimize their performance at scale - Implement and maintain observability solutions for complex distributed systems - Mentor junior engineers and contribute to building a culture of engineering excellence - Partner with product managers and other teams to define technical roadmaps and deliver customer-facing features A day in the life Cosmos is the foundational media and intelligence platform behind Amazon Photos, and this is a role where the code you write is felt by hundreds of millions of customers the same day it ships. You own the path that turns a raw photo into the crisp preview a customer sees the instant they open the app, serving images across every surface at up to 30,000 TPS, which makes it one of the most critical and most public experiences in the product, where performance, resilience, and graceful degradation are the job rather than nice-to-haves. You also build the computer vision and AI backbone that makes a customer's library feel intelligent, understanding what is in a photo and how good it is, picking out the best shots, recognizing people and the things that matter, spotting screenshots and text-heavy images, sharpening and enhancing older or lower quality pictures, and surfacing similar photos, so the app can curate Memories, power search, and guide editing and curation on the customer's behalf, all on a platform designed so a new capability can be added quickly rather than rebuilt from scratch, and you own the experiences that keep memories safe by making sure deletion and recovery happen correctly with the trust and privacy guarantees customers expect. About the team We operate at massive scale, processing billions of customer photos while maintaining high availability and low latency. Our team leverages technologies in distributed systems, computer vision, and machine learning to create innovative solutions that delight customers. We're building the future of how people interact with their memories, using AI to create more personalized and meaningful experiences.