Natural Language Processing
Your browser doesn't support speech here. You can still read everything, or try Chrome, Edge or Safari.

← All courses

Course 7 of 10 · Intermediate → Advanced

Natural Language Processing

Text, embeddings and LLMs. Tokenisation, embeddings, text classification and NER, text generation and evaluation, fine-tuning LLMs with SFT, LoRA and preference tuning.

How this course works

Welcome

  • Level: Intermediate → Advanced. You need: Deep Learning.
  • Before you start: the Deep Learning course, at least up to transformers.
  • Its game season follows the language projects of the Applied AI team.
  • Pick a topic from the sidebar; they go in order. Press ▶ Play this topic to hear a lesson read aloud, or 🔊 Explain on one section.
  • Topics climb from Beginner to Intermediate · Maths, Advanced and Code; use the Show chips to hide levels you don't need. Most have a Watch it animation, Try it demos and runnable code examples, and every topic ends with a quiz: answer it to mark the topic done ✓.
  • Stuck? Press 💬 Ask the AI tutor: a small free AI model that runs on your own device and answers from the course.
  • Progress is saved in this browser and shared across all the courses on this site.

Learning path: 🐍 Python → 🗄️ SQL & Databases → ∑ Maths for AI → 📈 Machine Learning → 🧠 Deep Learning → 👁️ Computer Vision → 💬 Natural Language Processing → ✨ Generative AI & LLM Apps → 🚀 Flask & Deployment → 🔬 Becoming an AI Researcher

NLP

1. NLP: Start HereWhat natural language processing covers today, the tasks and tools a junior NLP engineer uses, and how this track connects to the LLM topic and its game season.2. Text as Data: Normalisation & TokenisationFrom characters to tokens: normalisation, word and subword tokenisers (BPE, WordPiece, SentencePiece), vocabularies and special tokens, and why tokenisation affects cost, languages and bugs.3. Word & Sentence EmbeddingsMeaning as vectors: the distributional idea, word2vec and co-occurrence embeddings, analogies and bias, contextual embeddings, and sentence embeddings for search and clustering.4. Text Classification & Sequence LabellingBaselines first (TF-IDF + linear models), fine-tuning encoders, imbalanced and multi-label text, named-entity recognition with BIO tags and span-level F1, and when to use an LLM instead.5. 🎮 Milestone N1: The patient-message inboxRiverside's patient portal receives thousands of messages a week. Build the routing baseline with a safety threshold for urgent messages, measure the name-redaction model properly, learn embeddings from the inbox itself, and decide when an LLM is worth it. Includes 3 hands-on Python coding tasks (NumPy, scikit-learn), checked by automated tests.6. Language Models, Decoding & Evaluating Generated TextWhat a language model computes, perplexity, decoding strategies (greedy, beam, temperature, top-k, top-p), summarisation and translation, and how to evaluate generated text: overlap metrics, embedding metrics, LLM judges and humans.7. How LLMs Are Trained, Adapted & AlignedPretraining data and scaling laws, instruction tuning, preference alignment (RLHF and DPO), parameter-efficient fine-tuning, evaluation and contamination, safety, and how to choose between prompting, RAG and fine-tuning.8. 🎮 Milestone N2: Summaries clinicians can trustDoctors want draft summaries of long consultation notes. Choose decoding settings, evaluate summaries beyond ROUGE, align the summariser with clinicians' preferences using DPO, and decide how drafts are reviewed before anything enters a record. Includes 3 hands-on Python coding tasks (NumPy), checked by automated tests.9. ✅ Checkpoint: NLP trackA mixed quiz on the whole NLP track. Pass mark: 7 of 9 on the first try.
When you finish

Next course

Ready