LLM
Modern ML Practice in 2026
This is a short snapshot of practical machine learning work as of 2026. The biggest shift is that many projects now start from pretrained foundation models, but the hard work is...
Read noteNotes
Field notes from the ML workbench.
Short reviews, implementation reminders, and practical summaries for machine learning engineering.
L1
Regularization
Regularization is any training choice that helps a model generalize instead of only memorizing the training set. It can be a penalty in the loss, noise during training, constrai...
Read note One-hot Vectors
Embeddings
Machine learning models do not understand raw text directly. They need text to be converted into numeric vectors. An embedding is a learned vector representation for a token, wo...
Read note python
Python in Detail
This note reviews Python details that matter in day-to-day ML engineering: the data model, dictionaries, memory management, and concurrency choices.
Read note SGD
Optimizers
Optimizers update model parameters using gradients. The optimizer matters, but it is only one part of the recipe: initialization, normalization, batch size, learning-rate schedu...
Read note BatchNorm
ML: Normalization
Normalization makes optimization easier by controlling the scale and distribution of activations, features, weights, or gradients. The right normalization depends on the archite...
Read note Decision Tree
Ensemble Methods
Ensembles combine multiple models to improve generalization. The main idea is to reduce variance, bias, or both by making predictions from many weaker learners.
Read note sigmoid
Activation Functions
Activation functions introduce nonlinearity. Without them, a deep network is still equivalent to one linear transformation.
Read note