AWS Atlas

Amazon was founded over offering customers low prices across our wide selection of services. When Duolingo looked to optimize their cloud investments, they experienced this firsthand through their Technical Account Manager (TAM). In this session, Duolingo tech leadership and their TAM will share their journey: from the TAM’s approach to identifying cost savings to Duolingo’s experience implementing AWS Graviton processors.

Foundation models are able to solve a wide array of tasks, however to solve most tasks effectively, these models require some form of customization. In this session, you’ll learn about how you can get started with advanced customization of foundation models using SageMaker AI, while optimizing cost and performance.

AWS storage services can act as a powerful, cost-effective foundation for the most demanding high-performance workload, from genomics analysis to media rendering and real-time analytics. In this session, explore storage options that are suitable for numerous architectures, data types, and organizational requirements and that can accelerate your most performance-critical applications to speed up time to insights and time to market.

Explore Jefferies Equities' enterprise deployment of a generative AI trading assistant using Amazon Bedrock's Strands Agents framework. This lightning talk will showcase Jefferies’ technical architecture leveraging the Claude foundation model, Model Context Protocol (MCP) tools, and Amazon Bedrock Knowledge Bases for processing real-time trade and transactions data. We’ll demonstrate our serverless event-driven architecture for achieving lower latency response at scale.

Learn how Tempo Labs migrated production generative AI workloads from OpenAI to Amazon Bedrock, improving scalability, model choice, and security. Through AWS's Migration Acceleration Program (MAP) and generative AI Innovation Center, Tempo built a code-first UI/UX collaboration platform that enables product managers, designers, and engineers to collaborate directly in code—replacing image-based design workflows.

Operating in a multicloud environment can introduce operational complexity. In this session, learn to streamline operations with AWS Systems Manager, which allows for easier instance management across all environments. Gain insights into performance with Amazon CloudWatch and Amazon Managed Grafana, which delivers a unified dashboard of metrics and logs from any data source.

Data continues to be one of the world’s most valuable resources. Data protection and resilience are critical to business success, helping to keep workflows running and allow business recovery from data loss events such as unplanned outages, natural disasters, and ransomware.

Pinterest's management of petabyte-scale data lakes serving 500M+ users showcases innovative solutions for critical data challenges using Apache Iceberg on Amazon S3. You’ll learn practical strategies to optimize storage costs, enhance query performance, and maintaining data integrity at massive scale. See how Pinterest achieved breakthrough results, including significant error reduction through user agent-based access control and hash-based location patterns.

Traditional SIEM architectures tightly coupled data storage and threat detection, often leading to high costs, slower queries, and limited flexibility. In this session, learn how to reimagine your SIEM design by decoupling detection from storage using Amazon S3. Discover how to keep AWS-native logs in S3 while centralizing additional telemetry, such as endpoint detection and response (EDR) events, in the same buckets.

Unlock the full potential of Karpenter to streamline Kubernetes cost optimization. Dive deep into advanced strategies beyond basic auto-scaling. Learn how to fine-tune Karpenter’s resource allocation, and use dynamic configuration strategies to adapt to real-world workloads. Discover best practices and see how multi-layer automation can complement Karpenter to drive even greater impact on performance, scalability, and Kubernetes efficiency.

Discover how Slickdeals transformed from a heuristics-based platform to an AI-powered deal discovery engine using Amazon SageMaker. Learn how this e-commerce leader, which has saved shoppers over $10 billion, implemented five critical ML use cases including content summarization, deal personalization, and intelligent recommendations. We'll dive into their journey of modernizing their search and personalization algorithms, achieving enhanced user experiences for millions of shoppers.

Amazon file storage services continue to deliver innovations to help enterprise customers and storage admins manage their file data more effectively. Dive into the latest capabilities designed to enhance file data accessibility and integration with modern cloud applications. Discover how shared file storage from Amazon EFS and Amazon FSx is evolving to meet the needs of compute-intensive and data-heavy workloads.