NVIDIA neom agent toolkit uses Amazon s3 vectors for agent memory
The post explains how to use Amazon S3 Vectors as a persistent memory layer in NVIDIA NeMo Agent Toolkit (NAT). It walks through creating a vector bucket, implementing a custom MemoryEditor plugin, and configuring NAT to use the plugin. The tutorial targets teams running NAT on Amazon EKS and using Amazon Titan Text Embeddings V2 for vector generation.
Key points
- Amazon S3 Vectors can serve as a persistent memory backend for NVIDIA NeMo Agent Toolkit on Amazon EKS.
- The tutorial requires NAT 1.6, Python 3.11/3.12, Amazon Titan Text Embeddings V2, and an existing EKS cluster.
- Using the automemoryagent wrapper, agents automatically capture and retrieve memory, reducing duplicate work and improving groundedness.
The article lists prerequisites: an AWS account with S3 Vectors and EKS permissions, an existing EKS cluster, NAT 1.6, Python 3.11/3.12, and the embedding model. It then details three steps: set up the S3 Vectors infrastructure, build the MemoryEditor plugin, and wire the plugin into NAT’s YAML config with the automemoryagent wrapper. It also covers IAM roles, Docker images, and Kubernetes manifests.
The author discusses responsible AI considerations, such as retention policies, PII handling, and least‑privilege IAM scopes. It ends with a multi‑agent investment research use case that shows how shared memory reduces duplicate work and improves groundedness. The tutorial concludes with cleanup instructions to avoid ongoing charges.
Build agent memory with NVIDIA NeMo Agent Toolkit and Amazon S3 Vectors
AWS Machine Learning Blog · 1 October 2026
Loading the full article…
This text was published by AWS Machine Learning Blog and written by Venkata Sistla. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Agents & Tools
All →- Amazon releases Strands Decider 2B, a Jev-inspired open-source decision model · 2 src
- Decagon launches Voice 3 agent with Chord speech model for multilingual calls · 2 src
- Nvidia researchers improve AI agent reliability with a judging model · 1 src
- OpenAI launches dots agent to handle ongoing tasks without human input · 2 src
- Anthropic deploys Claude buying agent to handle inbound sales · 2 src
Comments
via GitHub Discussions