ML frameworks
| # | File | Covers |
|---|---|---|
| 01 | 01_what_is_pytorch.md | PyTorch fundamentals |
| 02 | 02_what_is_tensorflow.md | TensorFlow fundamentals |
| 03 | 03_pytorch_vs_tensorflow.md | the comparison |
| 04 | 04_pytorch_interview.md | PyTorch questions |
| 05 | 05_tensorflow_interview.md | TensorFlow questions |
| 06 | 06_choosing_a_framework.md | the 2026 map — scikit-learn, gradient boosting, JAX, Optuna |
The practical answer: scikit-learn is still the default for classical modelling and clean pipelines, gradient boosting still dominates structured data, and PyTorch is the default for deep learning because it’s what research and the entire LLM ecosystem target. You rarely choose — you use what your pretrained checkpoint was published in.