Agentic AI Course Development

This repository contains independently runnable Agentic AI teaching examples. Each example focuses on one clear concept and avoids unnecessary external services and API keys where possible.

Environment

• Python 3.12
• Windows PowerShell

Create and activate a virtual environment from the repository root:

py -3.12 -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install -U pip
python -m pip install -r requirements.txt

To run an LLM-backed example, copy .env.example to .env and add your OPENAI_API_KEY. Git does not commit .env. The example limits every agent response to 120 tokens to control API usage.

Run the first example:

python examples\langgraph_multi_agent\order_support.py

Examples

• langgraph_multi_agent: use customer-support ticket routing to understand coordination, specialist roles, and shared state.

Real LLM evaluation

After activating the virtual environment and configuring .env, run:

python scripts\run_eval.py

The runner calls a real LLM for the fixed cases in evals/order_support_cases.jsonl and scores routing, execution traces, required facts, and prohibited facts. Reports are written locally to evals/results/ and are not committed to Git.

Project files

• requirements.txt: third-party packages required to run the teaching examples.
• requirements-dev.txt: development tooling; currently contains ruff.
• .python-version: declares Python 3.12 for tools such as pyenv and uv; Windows users can still use py -3.12.
• pyproject.toml: shared tool configuration; currently stores Ruff code-style rules.
• .gitignore: prevents local virtual environments, caches, evaluation reports, and secrets from being committed.

Commit the configuration files and source code. Do not commit locally generated .venv/ or __pycache__/ directories.

Development checks

After installing development tooling, run:

python -m pip install -r requirements-dev.txt
python -m ruff check .