NLP
Natural language processing: tokenization, embeddings, classification, NER, sequence models, BERT and transformers.
What Is NLP?
The field, the tasks, and the deep-learning revolution that reshaped how machines read language.
Text Representation: From Tokens to Vectors
Tokenization, bag-of-words and TF-IDF — how raw text becomes numbers a model can use.
Text Preprocessing
Lowercasing, stop words, stemming and lemmatization — clean text without destroying meaning.
Regular Expressions for Text
Pattern matching for extraction, validation and cleaning — the precision tool in every NLP workflow.
N-grams & Language Models
Model word sequences with Markov assumptions, estimate probabilities, and generate text the old way.
Word Embeddings
Dense vectors that capture meaning: word2vec, GloVe, and using embeddings in models.
Text Classification
Spam, topics, intent: logistic regression and linear models on TF-IDF — fast and often enough.
Sentiment Analysis
Detect opinion and emotion in text — from lexicon scores to fine-tuned transformers.
Named Entity Recognition
Find people, places, organizations and dates — the extraction workhorse of document AI.
Part-of-Speech Tagging
Tag every token with its grammatical role — the classic sequence-labeling task.
Topic Modeling with LDA
Discover latent themes in a corpus with Latent Dirichlet Allocation.
Sequence Models for Text
Process text as a sequence with RNNs — the pre-transformer era of deep NLP.
LSTMs for Text
Gated memory for language: LSTMs and GRUs that actually hold context over long sequences.
Seq2Seq Models & Machine Translation
Encoder-decoder architectures that map one sequence to another — the basis of translation.
Attention & Transformers for NLP
Self-attention makes every token reach every other — the shift that unlocked modern NLP.
BERT & Pretrained Language Models
Masked language modeling and the pretrain-then-fine-tune recipe that powers NLP.
Fine-Tuning Transformers
Adapt a pretrained model to your task with the Hugging Face Trainer — clean and reproducible.
Question Answering
Extractive and generative QA — from span prediction to retrieval-augmented answers.
Text Summarization
Extractive and abstractive summaries with pretrained sequence-to-sequence models.
NLP Evaluation Metrics
Perplexity, BLEU, ROUGE and human judgment — measuring language models honestly.
NLP Roadmap
Synthesize the course, pick a specialization, and connect NLP to generative AI and LLM engineering.