// EXTRAS — companion modules
Beyond the main path.
Each main lesson keeps the FE-bridge framing. These extras drop the analogy and go deeper: math you derive by hand, classical ML you should reach for first, ethics you must consider before shipping, Python you read when fast.ai or Karpathy is the right next step.
classical ml
Logistic regression, decision trees, k-NN, ensembles. When NOT to reach for deep learning.
ethics responsibility
Dataset bias, fairness metrics (and which ones to pick), interpretability, responsible deployment.
- 25 min · beginnerDataset Bias: Garbage In, Discrimination Out
- 30 min · intermediateFairness Metrics: Defining 'Fair'
- 25 min · intermediateInterpretability: What Did the Model Actually Learn?
- 25 min · intermediateResponsible Deployment: Who Gets Hurt If You're Wrong?
- 30 min · intermediateBuilding an Ethics Checklist for Your ML Projects
math deep dive
Partial derivatives, chain rule for backprop, optimization landscape, FFT, KL divergence, Bayesian updates. Verified against tf.grad.
python bridge
Python for JS devs, NumPy ↔ TypedArray, PyTorch ↔ TFJS, pandas, reading ML papers.
- 20 min · beginnerPython in 15 Minutes for JavaScript Developers
- 25 min · beginnerNumPy = TypedArrays + Array Methods
- 30 min · intermediatePyTorch = TensorFlow.js (Different Syntax, Same Ideas)
- 25 min · intermediatepandas = Array.map/filter/reduce on Tabular Data
- 30 min · intermediateReading ML Papers and PyTorch Repos