RAG on AWS usign Amazon Bedrock, Amazon Kendra, Amazon Lambda Function, Amazon S3 and Amazon DynamoDB
π»πͺπ¨π± Dev.to Linkedin GitHub Twitter Instagram Youtube Linktr
In this repository you will find a use cases of RAG on AWS using CDK.
Additionally, you will find a notebook where you can run the agent localy.
- How to create a vectordb using Amazon Bedrock - Titan Embeddings, chroma and langchain.
- How to store the vectordb in an Amazon S3 bucket.
- How to build an Agent with memory capable of following a more fluid conversation (learn more about using memories with agents here)
- How to query vectordb stored in s3 bucket through a fluid conversation with langchain agent employed an LLM.
- How to query a Amazon Dynamodb table through a fluid conversation with langchain agent employed an LLM.
- How to put item to a Amazon Dynamodb table through a fluid conversation with langchain agent employed an LLM.
β AWS Level: Intermediate - 200
π° Cost to complete:
β Prerequisites:
This is a multilingual (limited to the LLM you use) agent is a spa assistant specialized in massages (with some data that I found on Google π ... it is not very reliable), recommends types of massages based on the client's comments, schedules appointments and checks their status.
β Clone the repo
git clone https://github.com/elizabethfuentes12/aws-qa-agent-with-bedrock-kendra-and-memory.git
β Create The DBvector Following The Code In This Notebook: Here
The dbvector will be saved within the CDK stack /vectordb, which will create an S3 bucket with the dbvector stored in it.
β Go to:
cd re-invent-agent
β Create The Virtual Environment: by following the steps in the README
python3 -m venv .venv
source .venv/bin/activate
for windows:
.venv\Scripts\activate.bat
β Install The Requirements:
pip install -r requirements.txt0
β Synthesize The Cloudformation Template With The Following Command:
cdk synth
β π The Deployment:
cdk deploy
Play with the agent and improve the prompt, remember that he has memory storage and you can have a fluid conversation with it.
π¨ Did you like this blog? π©π»βπ» Do you have comments?π€ tell me everything here
π»πͺπ¨π± Dev.to Linkedin GitHub Twitter Instagram Youtube Linktr
See CONTRIBUTING for more information.
This library is licensed under the MIT-0 License. See the LICENSE file.