About This Course

A 12-week NLP course for undergraduate students, bridging statistical language models and modern deep learning.

Course Structure

Part 1: Foundations (1-3)

N-grams, Word2Vec, RNN/LSTM

Part 2: Core (4-6)

Seq2Seq, Transformers, BERT

Part 3: Advanced (7-9)

Scaling, Tokenization, Decoding

Part 4: Applications (10-12)

Fine-tuning, Efficiency, Ethics

Learning Objectives

  • Implement n-gram models with smoothing techniques
  • Build word embeddings (Word2Vec, GloVe)
  • Design RNNs and LSTMs for sequence modeling
  • Implement attention and transformer architectures
  • Fine-tune BERT/GPT for downstream tasks
  • Apply LoRA, quantization, and distillation
  • Evaluate models and detect biases

Technology Stack

Python 3.8+ PyTorch Hugging Face Jupyter NumPy Matplotlib

Prerequisites

  • Python programming proficiency
  • Linear algebra basics (vectors, matrices)
  • Probability and statistics fundamentals
  • Intro ML concepts (gradient descent)

Resources

GitHub Download All PDFs Moodle HS25 QuantLet

Attribution

Developed by Digital AI Finance with QuantLet. All materials open-source.