LangChain and Vector Databases in Production — Course Notebooks

Hands-on notebooks for learning LangChain and production-oriented LLM workflows. This repo favors a private, local-first setup using Ollama with the qwen3:4b model — no OpenAI API keys required.

What's Inside

• Module 01: Foundations and guided walkthroughs.
  • Notebooks: see LLM-Systems-full/Module-01 - LangChain 101 Zero to Hero/LangChain_Complete_Guide.ipynb
• Module 02: LLMs and LangChain with Ollama (local models).
  • Notebooks:
    • LLM-Systems-full/Module-02 - LLMs and LangChain/Intro_to_LLMs.ipynb
    • LLM-Systems-full/Module-02 - LLMs and LangChain/Understanding_Tokens.ipynb

Key topics covered:

• LLM basics, prompting, and chaining with LangChain
• Chunking strategies (character and recursive) and classic split-process-combine
• Practical QA patterns (single/multi-question, batch processing)
• Tokenization concepts and demos (Hugging Face tokenizers), token budgeting

Prerequisites

• Python 3.10+
• Git
• Jupyter (Notebook or Lab)
• Ollama installed and running locally: https://ollama.com
• Model pulled locally (consistent with notebooks):

ollama pull qwen3:4b

If you prefer a different local model, set OLLAMA_MODEL in .env and adjust the notebooks accordingly.

Quick Start

1. Clone the repo

git clone https://github.com/williskipsjr/LangChain-and-Vector-DB.git
cd "LangChain & Vector Databases in Production"

2. Create a virtual environment and install deps

# Windows (PowerShell)
python -m venv .venv
. .venv\Scripts\Activate.ps1

# Install dependencies
pip install -r requirements.txt

3. Configure environment variables

# Copy the example and edit locally as needed
Copy-Item .env.example .env

Environment variables used by the notebooks:

• OLLAMA_BASE_URL: default http://localhost:11434
• OLLAMA_MODEL: default qwen3:4b
• LANGCHAIN_TRACING_V2, LANGCHAIN_ENDPOINT, LANGCHAIN_API_KEY: optional if you enable LangSmith/telemetry
• HF_TOKEN: optional for private models on Hugging Face (tokenizers in notebooks do not require this)

4. Launch Jupyter

jupyter lab
# or
jupyter notebook

Open the notebooks under Module 01 and Module 02 and run cells top-to-bottom. Ensure Ollama is running before executing any LLM cells.

Notes and Tips

• This repo avoids cloud keys by default. If you enable external providers, never commit secrets — .env is ignored; use .env.example to document variables.
• If you switch models, make sure to pull it with ollama pull <model> and update OLLAMA_MODEL.
• On Windows, large model downloads may take several minutes; keep the terminal open.

Troubleshooting

• Push protection/secrets: GitHub may block pushes that include credentials. Remove the secret and amend/re-push.
• Jupyter kernels: If the Python kernel isn’t found, reactivate the venv and reinstall Jupyter.
• tiktoken wheels: If installation fails, upgrade pip (python -m pip install -U pip) and retry.

Acknowledgements

• LangChain community packages
• Ollama for local model serving
• Hugging Face tokenizers for the token demos