Improving search quality autonomously with OpenSearch

Iteratively improving search quality through repeated experimentation has historically required specialized teams. As a result, organizations could apply this process to only a limited number of applications. Agentic applications increase the demand for high-quality search because their ability to complete tasks reliably often depends on the context they retrieve. In this talk, we present an agentic optimization loop that measures search quality based on agent telemetry, builds hypotheses for improvements, runs experiments, evaluates the results, and chooses the next iteration. We cover the technical foundations and platform capabilities the loop requires, along with implementation options for OpenSearch. You walk away with architecture patterns you can deploy to your own environment.

Search in the agentic era

AI-assisted search has improved information retrieval, moving from word-by-word matching to understanding abstract queries, context, and different query modalities. This evolution is far from over. Let's dive into the capabilities that exist today, using real-world examples, and explore what the future might hold.

AI-powered observability using OpenSearch

Do you have your observability story figured out? Even if you do, chances are that during a production incident, finding the root cause of the issue still takes a long time. In this talk, we'll explore three AI-powered methods that you can apply today to speed up troubleshooting and fixing issues using OpenSearch: natural-language query generation, semantic log search, and Model Context Protocol (MCP).

Vector search at scale: A deep dive into modern vector search techniques on OpenSearch

Comparing vectors is a fundamental capability in modern generative AI applications, enabling RAG and hybrid search use cases among others. Yet the details of how vectors are stored, retrieved, and compared by search engines are not widely understood. Modern vector search techniques allow for reduced memory and storage footprint and lower latency than naive exact comparisons, providing a better user experience at lower infrastructure cost. Join this session for a deep dive on modern vector search techniques and implementation in OpenSearch.

The Evolution of Search: From Keywords to Vectors and Beyond with OpenSearch

Vector search is all the hype, but how do you combine and integrate it with classic search techniques? In this session, we take a tour through the evolution of search, from simple string matching to lexical search to distributed search systems, semantic and AI-powered approaches. You learn what to keep in mind when building your search product or feature, and how OpenSearch can help you deliver.

OpenSearch Under Pressure: We Break It, You Fix It

Troubleshooting distributed systems takes practice—and OpenSearch is no exception. In this hands-on workshop, you'll investigate and resolve real-world issues in Amazon OpenSearch Service environments. From slow queries and indexing failures to shard imbalances and cluster instability, you'll perform root-cause analysis and remediate the issues. Strengthen your skills in diagnosing problems, applying fixes, and restoring performance. Whether you're an OpenSearch beginner or an experienced user, this session will sharpen your operational instincts and prepare you for production challenges. Get ready to get your hands dirty—we’ll break it, and you’ll fix it.

This is a hands-on workshop, you need your laptop to participate.

Building AI Agents with Amazon Bedrock AgentCore

Building production-grade generative AI applications is hard. Security, observability, scalability, and reliability all need to be designed and built besides orchestrating prompts. Amazon Bedrock AgentCore is a new set of AWS features that simplifies this heavy lifting and integrates with your favorite open-source framework. In this session, we explore AgentCore in depth and walk through a demo application running on AgentCore.