AWS Atlas

Data engineers face critical time sinks: writing code to build analytics pipelines from scratch and upgrading Apache Spark versions. In this lightning talk, discover how AWS is addressing both challenges with AI agents that accelerate development cycles. Learn how the Amazon SageMaker Data Agent transforms natural language instructions into executable SQL and Python code within SageMaker notebooks, maintaining full context awareness of your data sources and schemas.

How does AWS maintain operational excellence at a global scale? We'll dive deep into our operational best practices, culture, and learnings we've developed to operate large-scale services. We will also demonstrate some of the most-loved features in Amazon CloudWatch by AWS operators that you can readily deploy to improve operational efficiency and reduce Mean Time To Resolve (MTTR).

Organizations operating in multicloud environments face challenges with distributed data security, identity management, and continuous monitoring. In this session, we will explore approaches to address multicloud security, aligned with NIST Cybersecurity Framework (CSF), an industry standard, focused on protect, detect, and respond functions. We will dive deep into security visibility via centralized logging, encryption and privacy enforcement, and centralized federation.

Are you looking to customize foundation models and deploy AI agents at scale? Learn how to leverage Amazon SageMaker AI to build performant agentic workflows with customized open-weight models. This session covers the end-to-end journey: use Amazon SageMaker AI for model customization, track experiments with managed MLflow, establish governance with Amazon SageMaker AI Model Registry, and deploy optimized models using SGLang on Amazon SageMaker AI Inference for low-latency agent applications.

This breakout session dives into licensing commercial software on AWS, with a focus on Microsoft workloads. Join our Microsoft Optimization and Licensing Assessment (OLA) specialists to explore the best practices for navigating licensing complexities in the cloud while maximizing cost-efficiency. Learn about fully-funded OLA offerings for on-premises workloads, cloud to AWS migrations, existing EC2 deployments, and the new self-serve AWS Transform Assessments.

Navigate the intersection of AI and security culture in this practical session. Learn how security teams can effectively embrace AI innovation through incremental steps and validation techniques. Using real-world examples, we'll demonstrate how security practitioners can adapt their skills to AI challenges regardless of their level of specialized expertise and share strategies for building security-aware AI practices.

Learn how Amazon S3 Metadata automatically generates rich metadata for your S3 objects in near real-time. We’ll show how S3 Metadata creates a queryable metadata layer that transforms how you discover and manage your data. Watch as we explore and filter objects based on attributes like object creation time and storage class to streamline data preparation for analytics, real-time inference, building automated storage dashboards, and more.

Amazon EBS General-Purpose (gp3) volumes are the foundation for a wide range of workloads, including dev/test environments, low-latency databases, and analytics pipelines. In this session, learn how to leverage the latest gp3 enhancements to unlock higher performance requirements of your applications. Dive into the decoupled performance model of gp3 and how to dynamically tune throughput and IOPS based on workload patterns to achieve optimal results.

Join Samsung Cloud to learn how they successfully implemented a massive-scale table optimization, handling 1.2 PB on Amazon DynamoDB with zero downtime. Follow the innovative journey behind Samsung Internet's tab sync feature, showcasing how the team developed an in-house solution to handle complex data transfer achieving substantial cost savings without third-party dependencies.

Legacy data warehouses often slow innovation with outdated code, silos, and costly migrations. In this session, we’ll share proven best practices to modernize with confidence. Learn how to profile and assess source systems, convert legacy SQL at scale with automation, and validate data integrity to avoid surprises. We’ll cover an end-to-end process—from discovery through reconciliation—that reduces risk, improves predictability, and accelerates migration timelines by up to 50%.

As environments become more diverse and distributed, detecting and investigating security risks become increasingly complex. Discover how the new security analytics in Amazon OpenSearch Service streamlines security investigations and reduces costs by leveraging Amazon CloudWatch [Unified Data Store] for near-real time security analytics without full log indexing. Through demonstrations, learn how to create custom detection rules, implement correlation analytics, and configure notifications.

Traditional Kubernetes scaling on Amazon EKS often leads to a tradeoff between performance and cost. Reactive autoscaling based on utilization metrics can’t keep pace with dynamic workloads, resulting in wasted resources or degraded performance. This session introduces a proactive approach to resource management using workload-aware policies that distinguish between critical services and batch workloads.