01
Foundations and tokenization
Data preparation, tokenization, and the design choices that shape a language model before training begins.
Courses6 links
LLMIn progress
Large Language Models
From data and tokenization to training, evaluation, and serving
A practical learning track for understanding how modern language models are built, measured, and operated.
Learning map
The course index connects public course material with my own notes and experiments. It is designed to make the path from official source to searchable learning asset explicit.
Resource index
Course materials
Official
Stanford CS336: Language Modeling from Scratch
The current official course website, covering the full language-model development lifecycle.
Slides
CS336 course materials
Slides, schedule, assignments, and supporting material are maintained with the official course site.
Notes
Model-building notes
Personal notes and experiments for building compact language models.
Transcripts
CS336 notes and transcripts
The working collection for lecture notes and transcript-style explanations.
Blog
LLM systems writing
Follow-up writing on scaling, evaluation, and the engineering details behind LLM systems.
Notes
Stanford CS336 2026
Course resources and published material on the learning site.
DataHub workflow
01
Official course material
02
Slides and lecture notes
03
Transcript and explanation
04
Runnable experiments
05
Technical blog
Syllabus
02
Transformer architecture and optimization
The model, objective, optimizer, and systems constraints that determine training behavior.
03
Scaling, evaluation, and deployment
How to reason about scaling laws, reliable evaluation, inference efficiency, and serving.