How much Python do I actually need to build with AI?

Surprisingly little. To call an AI API you need variables, dictionaries, functions, a way to send an HTTP request, and a way to read JSON responses. This crash course covers exactly that core, skipping the topics you can learn later as you need them.

Why do I need a virtual environment?

A virtual environment keeps each project's packages separate so one project's dependencies cannot break another's. Create it once per project with python -m venv .venv, activate it, and install packages inside it. It is a one-time habit that prevents most beginner package conflicts.

What is the difference between a list and a dictionary?

A list is an ordered collection accessed by position, like items = ["a", "b"]. A dictionary maps keys to values, like user = {"name": "Ada"}, accessed by key. API responses are almost always nested dictionaries and lists, so you read them by chaining keys and indexes together.

Do I need to know maths or machine learning theory first?

No. To use AI APIs you are calling a service someone else trained, so you need programming basics, not the underlying maths. Machine-learning theory matters only if you want to train or fine-tune models yourself, which is a separate, later step.

Python for AI: A Beginners Crash Course

Why a focused crash course

Most Python tutorials spend weeks on material you do not need to start building with AI. The truth is that calling a modern AI API requires a small slice of the language: a few data types, loops, functions, and the ability to send an HTTP request and read a JSON response. This crash course teaches that slice directly, so you can go from zero to a working API call quickly and learn the rest on demand. Run every example yourself — reading code is not the same as writing it.

Setting up Python the right way

Install Python 3.10 or newer from python.org, then for each project create an isolated environment so your packages do not collide:

python -m venv .venv
source .venv/bin/activate    # on Windows: .venv\Scripts\activate
pip install openai

The virtual environment is a folder that holds this project’s packages only. Activate it whenever you work on the project. This single habit prevents the most common beginner frustration — packages from one project breaking another.

The data types you will actually use

Three structures cover almost everything in AI work. Variables hold values (model = "gpt-4o-mini"). Lists hold ordered collections (messages = [m1, m2]). Dictionaries map keys to values ({"role": "user", "content": "Hi"}). API requests and responses are built entirely from nested dictionaries and lists, so reading a response means chaining access: data["choices"][0]["message"]["content"]. Spend your first hour becoming fluent in those three and you can read most AI code.

Loops, functions, and JSON

A loop repeats work — iterating over a list of prompts, for example. A function packages reusable logic so you can call ask(prompt) instead of copying request code everywhere. JSON is the text format APIs speak; Python’s json module and the requests/openai libraries convert between JSON text and Python dictionaries automatically. Knowing how to loop over results and wrap your call in a function is enough to build a real script.

Your first AI API call

With the openai package installed and your key set as an environment variable (export OPENAI_API_KEY=sk-...), a complete first program is short:

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
resp = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)
print(resp.choices[0].message.content)

Never hard-code your key in the file — read it from the environment so you can share code safely. From here, everything else (loops over many prompts, saving results, building a web app) is a combination of the basics above. Use the lesson navigator below to step through each concept with its own runnable snippet.