Beginner AI Cloud CV Project📝

Simply Explained...

Project Metadata:

Project Plan:

This week, we're going to deploy an AWS Lambda function containing a Python script with the appropriate permissions to query a large language model that we're hosting in SageMaker. We'll also add contextualised data using Amazon S3, which the model can reference.

Why?

LLMs are powerful, but when deployed in real world scenarios, we often want them to understand and respond using our own relevant data. Maybe that's internal documents, logs, or domain-specific knowledge. This approach is sometimes called lightweight RAG, it helps us inject relevant, custom data into the model’s prompt without the complexity of a full on RAG (I recommend looking more into this part, but for the project you just need a basic understanding)

Let’s get to it…

Step One:

First, we need to deploy our Lambda function. It is recommended to do this via Terraform or the AWS CDK, but if you are more comfortable with the console, feel free to use that for now. Since most of my readers are just starting out, we will use the console for this guide.

Navigate to the Lambda console and select Create function.

Configure the Lambda function to use Python (choose the latest supported version, such as Python 3.12), and set the timeout to 1 minute.

This next part is really simple…