Natural Language Processing

Every NLP model is only as good as the tokens you feed it, so clean, consistent text prep is the foundation of the whole pipeline.

Bag-of-words and TF-IDF are simple, fast baselines that still beat many neural models on small datasets and teach the vectorization mindset.

Embeddings replace sparse one-hot vectors with dense vectors that capture meaning, enabling analogies and powering every modern NLP model.

Classification (spam, topic, sentiment) is the workhorse NLP task and the fastest way to turn text features into something useful.

RNNs/LSTMs introduce the sequential, stateful modeling that motivated attention and transformers, so understanding them explains what came next.

The transformer is the architecture behind virtually every modern NLP system, and self-attention is the key idea you must internalize.

Hugging Face is the de facto toolkit for applying pretrained transformers, letting you get state-of-the-art results in a few lines.

Subword tokenization is how transformers handle huge vocabularies and unseen words, and understanding it demystifies model inputs.

Fine-tuning adapts a powerful pretrained model to your specific task with modest data, and it is the core skill for delivering real NLP results.

A capstone forces you to integrate preprocessing, features, models, and evaluation into one deliverable you can show and defend.