Skip to content

Why Python?

Python shows up almost everywhere in modern software work. Web APIs run on FastAPI, Django, and Flask.

The data and scientific stacks (pandas, numpy, scipy, Jupyter) have been the default tools in their fields for years.

Most ML and AI work happens through Python: PyTorch, TensorFlow, the Hugging Face ecosystem, and nearly every model-serving library expose Python first. On the operations side, Python is the glue language for infrastructure scripts, automation, CI pipelines, and one-off tools that have to read a CSV and hit an API.

Thoughts on AI & Vibe Coding

Python is heavily represented in the training data of every major LLM, but unfortunately training models can be out of date.

Warning

A lot of LLMs will try to ship untyped code, not use the latest tools, or use libraries that are unmaintained or haven't been updated in three years.

Python code can compile and run even when it's wrong. Always check generated code.

The goal is build a foundation that lets you read AI output critically, accept what's correct, and reject what isn't.

Make sure your agents know how Python projects should be structured, what tools should be used, how to install packages, etc. I'll share a resource at the end of the course.

Shipping code you don't understand to a Production environment can get you in trouble. Write and read Python you understand, whether you typed it or a model did.