Course links

Course	Links	Role
Stanford CS224N, Spring 2024 — Christopher Manning, Natural Language Processing with Deep Learning	Course · Lecture playlist	Main course: word vectors, language models, pretraining, post-training, evaluation, reasoning and agents, machine translation, and parsing (chapters 2–3, 5, 9–16). Its four assignments cover word vectors, a neural dependency parser, attention-based translation, and pretraining and fine-tuning a small transformer.
Stanford CS336, Spring 2025 — Percy Liang and Tatsunori Hashimoto, Language Modeling from Scratch	Course · Lecture playlist · Lecture materials · Spring 2026 recordings	Main course: building a language model end to end, from tokenization, architectures, mixture of experts, and resource accounting to scaling laws, inference, evaluation, data, and alignment with reinforcement learning (chapters 1, 4, 6–8, 11–12, 14). Several lectures are executable Python files viewed in the course’s trace viewer. Its five assignments build a tokenizer and transformer, systems optimizations, a scaling-law study, a data pipeline, and post-training for mathematical reasoning.
CMU 11-711, Fall 2025 — Sean Welleck, Advanced Natural Language Processing	Course page with slides · Code	Supplement: in-context learning, fine-tuning, decoding and inference-time algorithms, retrieval, evaluation and experimental design, reinforcement learning for language models, and agents (chapters 5, 8–10, 12–14). The code repository has short notebooks for many lectures.

Code resources complement the courses:

Resource	Link	Role
Andrej Karpathy, Neural Networks: Zero to Hero	Playlist	Code-along lectures from a counting bigram model to a GPT, and a byte-level BPE tokenizer (chapters 1, 2, and 4).
nanoGPT and minbpe	nanoGPT · minbpe	Minimal, readable implementations of a GPT training loop and a BPE tokenizer (chapters 1 and 4).
CS336 assignments	Course	The five assignments with tests and handouts, the most thorough exercises for chapters 1, 4, 6–7, and 11–12.

The reading plan maps individual lectures, slides, and readings to the study sequence. Video links arranges the recordings by chapter, and the NLP overview maps the reading-plan topics to the chapters.