NLP — Foundations to Advanced
Language is the interface to intelligence. Master every layer.
29 lessons · 0 visual stories
0/29 complete · 0%
#LessonTypeLanguagesTimeOpen
01Text Processing — Tokenization, Stemming, LemmatizationLanguage is continuous. Models are discrete. Preprocessing is the bridge.BuildPython45 min↗ Official lesson02Bag of Words, TF-IDF, and Text RepresentationCount first, think later. TF-IDF still beats embeddings on well-defined tasks in 2026.BuildPython75 min↗ Official lesson03Word Embeddings — Word2Vec from ScratchA word is the company it keeps. Train a shallow net on that idea and geometry falls out.BuildPython75 min↗ Official lesson04GloVe, FastText, and Subword EmbeddingsWord2Vec trained one embedding per word. GloVe factorized the co-occurrence matrix. FastText embedded the pieces. BPE bridged to transformers.BuildPython45 min↗ Official lesson05Sentiment AnalysisThe canonical NLP task. Most of what you need to know about classical text classification shows up here.BuildPython75 min↗ Official lesson06Named Entity RecognitionPull the names out. Sounds easy until you deal with ambiguous boundaries, nested entities, and domain jargon.BuildPython75 min↗ Official lesson07POS Tagging and Syntactic ParsingGrammar was unfashionable for a while. Then every LLM pipeline needed to validate structured extraction, and it came back.BuildPython45 min↗ Official lesson08CNNs and RNNs for TextConvolutions learn n-grams. Recurrences remember. Both are superseded by attention. Both still matter on constrained hardware.BuildPython75 min↗ Official lesson09Sequence-to-Sequence ModelsTwo RNNs pretending to be a translator. The bottleneck they hit is the reason attention exists.BuildPython75 min↗ Official lesson10Attention Mechanism — The BreakthroughThe decoder stops squinting at a compressed summary and starts looking at the whole source. Everything after this is attention plus engineering.BuildPython45 min↗ Official lesson11Machine TranslationTranslation is the task that paid for NLP research for thirty years and keeps paying now.BuildPython75 min↗ Official lesson12Text SummarizationExtractive systems tell you what the document said. Abstractive systems tell you what the author meant. Different tasks, different pitfalls.BuildPython75 min↗ Official lesson13Question Answering SystemsThree systems shaped modern QA. Extractive found spans. Retrieval-augmented grounded them in documents. Generative produced answers. Every modern AI assistant is a mix of the three.BuildPython75 min↗ Official lesson14Information Retrieval and SearchBM25 is precise but brittle. Dense casts a wide net but misses keywords. Hybrid is the 2026 default. Everything else is tuning.BuildPython75 min↗ Official lesson15Topic Modeling — LDA and BERTopicLDA: documents are mixtures of topics, topics are distributions over words. BERTopic: documents cluster in embedding space, clusters are topics. Same goal, different decompositions.LearnPython45 min↗ Official lesson16Text Generation Before Transformers — N-gram Language ModelsIf a word is surprising, the model is bad. Perplexity makes surprise a number. Smoothing keeps it finite.BuildPython45 min↗ Official lesson17Chatbots — Rule-Based to Neural to LLM AgentsELIZA replied with pattern matches. DialogFlow mapped intents. GPT answered from weights. Claude runs tools and verifies. Each era solved the previous one's worst failure.LearnPython75 min↗ Official lesson18Multilingual NLPOne model, 100+ languages, zero training data for most of them. Cross-lingual transfer is the practical miracle of the 2020s.LearnPython45 min↗ Official lesson19Subword Tokenization — BPE, WordPiece, Unigram, SentencePieceWord tokenizers choke on unseen words. Character tokenizers blow up sequence length. Subword tokenizers split the difference. Every modern LLM ships on one.LearnPython60 min↗ Official lesson20Structured Outputs & Constrained DecodingAsk an LLM for JSON. Get JSON most of the time. In production, "most" is the problem. Constrained decoding turns "most" into "always" by editing the logits before sampling.BuildPython60 min↗ Official lesson21Natural Language Inference — Textual Entailment"t entails h" means a human reading t would conclude h is true. NLI is the task of predicting entailment / contradiction / neutral. Boring on the surface, load-bearing in production.LearnPython60 min↗ Official lesson22Embedding Models — The 2026 Deep DiveWord2Vec gave you a vector per word. Modern embedding models give you a vector per passage, cross-lingual, with sparse, dense, and multi-vector views, sized to fit your index. Pick wrong and your RAG retrieves the wrong thing.LearnPython60 min↗ Official lesson23Chunking Strategies for RAGChunking configuration influences retrieval quality as much as the choice of embedding model (Vectara NAACL 2025). Get chunking wrong and no amount of reranking saves you.BuildPython60 min↗ Official lesson24Coreference Resolution"She called him. He did not answer. The doctor was at lunch." Three references to two people and nobody is named. Coreference resolution figures out who is who.LearnPython60 min↗ Official lesson25Entity Linking & DisambiguationNER found "Paris." Entity linking decides: Paris, France? Paris Hilton? Paris, Texas? Paris (the Trojan prince)? Without linking, your knowledge graph stays ambiguous.BuildPython60 min↗ Official lesson26Relation Extraction & Knowledge Graph ConstructionNER found the entities. Entity linking anchored them. Relation extraction finds the edges between them. A knowledge graph is the sum of nodes, edges, and their provenance.BuildPython60 min↗ Official lesson27LLM Evaluation — RAGAS, DeepEval, G-EvalExact-match and F1 miss semantic equivalence. Human review does not scale. LLM-as-judge is the production answer — with enough calibration to trust the number.BuildPython75 min↗ Official lesson28Long-Context Evaluation — NIAH, RULER, LongBench, MRCRGemini 3 Pro advertises 10M tokens of context. At 1M tokens, 8-needle MRCR drops to 26.3%. Advertised ≠ usable. Long-context evaluation tells you the actual capacity of the model you are shipping on.LearnPython60 min↗ Official lesson29Dialogue State Tracking"I want a cheap restaurant in the north... actually make it moderate... and add Italian." Three turns, three state updates. DST keeps the slot-value dict in sync so the booking works.BuildPython75 min↗ Official lesson