DaaS / Products / Event-driven embedding inference pipeline

Event-driven embedding inference pipeline

Deploy an embedding model in OpenSearch to serve vector embeddings, then configure EventBridge to automatically deliver new document events (e.g., OSS uploads) to the inference API endpoint, creating an automated pipeline that embeds and indexes content as it arrives.

Products involved

Scenario

Deploy an embedding model in OpenSearch to serve vector embeddings, then configure EventBridge to automatically deliver new document events (e.g., OSS uploads) to the inference API endpoint, creating an automated pipeline that embeds and indexes content as it arrives.

How the products combine

  1. eb · eb-deliver-destinations — EventBridge — Deliver events to external destinations (API/OSS/MQTT)

    2. See eb/eb-deliver-destinations.

  2. opensearch · opensearch-deploy-model — OpenSearch — Deploy embedding model for inference

    3. See opensearch/opensearch-deploy-model.

Typical questions

FAQ

Q: How does the event-driven embedding pipeline automatically trigger vectorization on document upload? A: The pipeline automatically embeds and indexes content by configuring EventBridge to deliver new document events directly to an OpenSearch inference API endpoint. This architecture combines EventBridge destination delivery with a deployed OpenSearch embedding model to create a fully automated vectorization workflow.