Next Gen Agentic Search on AWS
summary
Checkout.com migrates from Elasticsearch 7.10.2 to Amazon OpenSearch Service using a staged approach: dual-write checker to validate parity, snapshot restore for historical data, read cutover followed by write cutover. The payments stack ingests via Kinesis/SNS/Lambda into OpenSearch and serves billions of transactions; migration causes P99 latency spikes from long-range queries which get resolved by query optimization and right-sizing. Plan moves to OpenSearch 3 for tiered storage, cost and agentic features — agentic framework with memory and vector/semantic search.
AI-written summary · about
chapters (4)
- 0:00Introduction to agentic search
- 5:58Checkout.com’s migration story
- 15:48Future of agentic search
- 25:37Q&A session
related talks
40mBuilding Intelligent Search Systems With Agentic RAG and OpenSearch
36mBuilding Next-gen Search With OpenSearch
06mSponsored Keynote: OpenSearch - See Everything: Open Observability for Agentic AI
38mAI-Ready Data Foundation Using Amazon OpenSearch
25mKeynote: OpenSearch in the Agentic Era
24mThe Aryn Journey - Growing a Gen AI Startup on OpenSearch and the Road Ahead
30mOpenSearch for GenAI Storage: AgentScope Studio Practice
47mBoosting AI With OpenSearch: Vector Stores, Similarity Search, and Retrieval-Augme...