// the find
NeuralNine/youtube-tutorials
A collection of the code I have written for my YouTube tutorials.
This is NeuralNine's grab-bag of companion code for his YouTube channel: well over 100 standalone folders, each a mini-project demonstrating one library or concept — Flask apps, ML notebooks, Docker examples, MCP servers, LoRA fine-tuning scripts, and more. It's for people who watch the videos and want the exact code to follow along, not a library or framework you'd depend on.
Impressive breadth and currency — folders for MCP servers, CrewAI, Axolotl LoRA fine-tuning, and Bifrost show the channel tracks whatever's current in the AI tooling space within weeks of release. Each topic is isolated in its own folder so you can clone one example without dragging in unrelated dependencies. Several projects go past toy-script level into full stacks (FastAPI/React in 'AI Stock Analysis Assistant', Django+Celery in 'Job SaaS'), which is more than most tutorial repos bother with.
There's no root README or index mapping folders to video titles, so with 100+ projects and names like 'main.py'/'main2.py'/'main3.py' repeated across folders, finding the one tied to a specific video is guesswork. Dependency management is inconsistent — some projects have pyproject.toml/uv.lock, others a bare requirements.txt, several have neither, so half of these won't run without reverse-engineering imports. Multiple folders commit live-looking '.env' files directly into the tree, which is a bad pattern to model for viewers even if the values are dummies. None of this is tested or maintained as a unit — it's a changelog of disconnected demos, so expect bit rot in the older folders as libraries move on.