Open-source course by Fast Batch
Welcome! This is an open-source course that walks you through building a smart home agent from scratch. This is a hands-on project that covers the full stack, from agent design and LLM integration to MLOps pipelines, monitoring, deployment infrastructure, and everything in between.
This is not a series with a fixed timeline. This is a long-term journey where we’ll build, break, and rebuild until we have something that actually works in production.
-
Build the agent architecture and reasoning logic
-
Integrate evaluation and testing strategies
-
Implement observability and monitoring tools
-
Secure your agent with safety guardrails
-
Design deployment pipelines and infrastructure
This repo accompanies the course published on Substack, where each module is explained with code walk-through and lessons learned.
1- Create a virtual environment and install the project dependencies
uv pip install -e .2- Create a .env file by coping the .env_example file and add the necessary API keys
3- Setup Opik for monitoring
opik configureYou will need to login and get the API key from your account 4- Run your agent
python run.py
This open-source series consists of 8 modules. This course is still under-development, thus more modules can be added.
The best way to follow along is to clone the repository and run the code. We purposefully setup different branches for different modules to make it easy to navigate through the code.
| Module | Substack | Description | Running the code |
|---|---|---|---|
| 0 | Introduction | Series announcement | - |
| 1 | Build your smart home agent | Implement a simple agent that controls devices | Demo |
| 2 | Personnalize your agent | Add a memory layer to personnalize your agent's response | Demo |
| 3 | Deploy your agent | Design the deployment architecture | Demo |
| 4 | Evaluate your agent | Evaluate if your agent if ready for production | Demo |
| 5 | Red team your agent | Evaluate your agent's robustness against attacks | - |
| 6 | Secure your agent | Implement safety guardrails | - |
| 7 | Monitor your agent | Add observability and monitoring pipelines | - |
| 8 | Scale up your agent | How to scale your deployment pipele | - |
This is a solo project for now, but PRs that improve clarity, fix bugs, or suggest useful utilities are welcome. If you’re building something similar, feel free to reach out or open an issue to discuss!