Building Agent Memory Using OpenSearch & GenAI Observability
September 19, 2025
2:00 PM - 4:00 PM (IST)
Virtual
Join us for an intensive technical session exploring the architecture and implementation of persistent agent memory systems using OpenSearch. This specialized workshop focuses on building sophisticated memory frameworks aligned with Amazon Bedrock's Agent Core Memory patterns, enabling AI agents to maintain context across interactions and learn from past experiences.
Through expert demonstrations and technical discussions, participants will learn how to design and implement memory persistence layers that enhance agent capabilities while maintaining performance at scale. Discover how to effectively leverage OpenSearch's capabilities to create memory systems that give your AI applications a significant competitive advantage through improved contextual understanding and response relevance.
As organizations increasingly deploy RAG systems and large language
models in production environments, the need for visibility into their
operation becomes critical. This session will also explore the unique
challenges of monitoring generative AI systems and provides practical
approaches for implementing metrics, logging, and evaluation frameworks
AI/ML engineers building
agent-based applications with Amazon Bedrock
Cloud architects designing memory persistence solutions for conversational AI
Software developers implementing stateful AI agents
Data engineers working with vector search and knowledge retrieval
Technical leaders evaluating memory options for generative AI systems
Product managers overseeing advanced conversational agent implementations
AWS practitioners looking to extend Bedrock agent capabilities
Introduction to Agent Memory Systems
OpenSearch as a Memory Store
Memory Implementation Patterns
Performance Optimization and Scaling
Fundamentals of GenAI Observability
Muhammad Ali is a Principal Analytics (APJ Tech Lead) at Amazon Web Services (AWS), based in Melbourne, Victoria, Australia. With nearly 7 years at AWS, he specializes in data analytics, artificial intelligence, and cloud technologies.
As a Principal OpenSearch Solutions Architect, he works closely with customers across Asia Pacific to design and scale complex search and log analytics systems. Ali is actively involved in the tech community, leading the Melbourne OpenSearch User Group meetup. and his daily collaboration with open-source maintainers and contributors ensures he stays at the forefront of the OpenSearch community.
Dhinesh Raja M is a Solutions Architect at AWS, committed to
collaborating closely with customers to offer guidance and technical support for their large-scale cloud initiatives on AWS. His current focus lies in guiding independent software vendors (ISVs) through their cloud adoption journey, with his area of depth being GenAI and AI/ML, aiming to drive strategic business outcomes.
Learn patterns for implementing memory systems that align perfectly with Amazon Bedrock agents and open source, enabling more coherent and contextually aware
AI conversations.
Understand how properly implemented memory systems dramatically improve agent performance by providing relevant historical context and learned information.
Learn how to implement observability systems that transform opaque AI behaviors into transparent, measurable processes you can confidently manage and improve.
Learn how to seamlessly connect your memory systems with existing AWS infrastructure, creating a unified technical ecosystem for your AI applications.
Understand how to implement efficient memory persistence that balances performance needs with operational costs, maximizing your ROI.