AWS examples for a makeathon

🚀 Check out this document Quick Start Guide

1. Login
2. Roles & Permissions
3. Access Keys
4. Sagemaker AI Platform & Jupyter Notebooks
5. S3 Storage

All TypeScript examples (Bedrock, S3, S3 Vectors, LangChain, RAG) are in the typescript/ folder with their own setup, docs, and README.

👉 Go to TypeScript README

Quick overview of what's inside:

Checkout the S3_Example.ipynb notebook. → You can run this on AWS Sagemaker Notebooks

Checkout the Bedrock_Example.ipynb notebook. → You can run this on AWS Sagemaker Notebooks

Make sure you never store access keys in a public location!

1. Create a virtual python environment python3 -m venv .venv
2. Activate the virtual environment source .venv/bin/activate
3. Install the required libraries pip install -r requirements.txt

Source: https://docs.python.org/3/library/venv.html

1. Create an AWS Access key Link
2. Create a copy of the .env.example file and name it .env
3. Store the Key ID and the Key Secret in the .env file

WARNING Make sure you NEVER add these keys to a public repository!

ATTENTION! When selecting a model in Bedrock it's now required their inference profile IDs instead of raw model IDs.

Example:

• Before: amazon.nova-pro-v1:0
• Now: eu.amazon.nova-pro-v1:0

Find all available inference profile IDs here: Supported Regions and models for inference profiles (always choose EU when available, otherwise global). When generating code with AI you might need to change this manually, as it's a recent change in AWS ;)

1. Make sure the environment is activated source .venv/bin/activate
2. Execute python bedrock_local.py

The output should look like this:

(.venv) user@host$ python bedrock_local.py

>>> Model Response: A 'hello world' program demonstrates the basic syntax of a programming language by outputting a simple greeting.

1. Make sure the environment is activated source .venv/bin/activate
2. Change the group name in the file s3_local.py to a valid group name
3. Execute python s3_local.py

The output should look like this:

(.venv) user@host$ python s3_local.py
Downloaded 's3://makeathontest/data/dummy_data.csv' and stored file in 'dummy_data.csv'

(.venv) user@host$ ls
dummy_data.csv

If you're using an AI coding assistant (Cursor, Windsurf, Claude Code, GitHub Copilot, etc.), you can give it direct access to the latest LangChain documentation via their MCP server. This means your assistant will give you accurate, up-to-date LangChain code instead of hallucinating outdated APIs.

MCP Server URL:

https://docs.langchain.com/mcp

Claude Code:

claude mcp add --transport http docs-langchain https://docs.langchain.com/mcp

Cursor / Windsurf — add to your MCP settings (.cursor/mcp.json or equivalent):

{
  "mcpServers": {
    "langchain-docs": {
      "type": "http",
      "url": "https://docs.langchain.com/mcp"
    }
  }
}

Once connected, your assistant can search LangChain, LangGraph, and LangSmith docs in real time. More details: docs.langchain.com/use-these-docs

• Always use eu. inference profile IDs for Bedrock models to keep data in EU regions
• S3 bucket names must be lowercase — only letters, numbers, and hyphens, globally unique
• SSO credentials expire after ~1 hour — re-copy from the portal when you get ExpiredTokenException