Hands-on Python notebooks accompanying the "From Vectors to Vibes" blog post series. Each week we take one section from the post, walk through the concept, and build it together in code.
The goal: give every developer at Apex Lab enough understanding of why AI works — not just how to use it — so you can engineer with it, not just prompt at it.
Requires Python 3.14.
./setup.shThis will:
- Create a
.venvvirtual environment with Python 3.14 - Install all dependencies from
requirements.txt - Register a Jupyter kernel named "AI Engineering (Python 3.14)"
Then either:
source .venv/bin/activate
jupyter notebook notebooks/Or open in VS Code / Cursor and select the "AI Engineering (Python 3.14)" kernel.
| # | Notebook | Section | Topics |
|---|---|---|---|
| 01 | 01-vectors-and-similarity.ipynb |
0.1 — Vectors | Scalars, vectors, matrices, tensors; 2D word space; Euclidean distance vs cosine similarity; vector arithmetic (king − man + woman ≈ queen); real embeddings with sentence-transformers; semantic search demo |
| 02 | 02-embeddings.ipynb |
0.2 — Embeddings | Word2Vec from scratch; BPE tokenization (from scratch + tiktoken); word vs sentence vs code embeddings; cross-language code search; dimension tradeoffs; mini RAG pipeline |
| 03 | 03-neural-networks.ipynb |
0.3 — Neural Networks | Interactive single neuron; activation functions (ReLU, sigmoid, tanh, softmax); interactive 2-layer network; training on MNIST; visualizing learned features per layer; parameters & scale |