Course

1. Home›
2. Course

Five modules take you from the foundations of neural networks through model mechanics, grounded systems, quality engineering, and controlled autonomy. Each lesson builds on the last; nothing is used before it is defined.

1. 

   Module 0基礎

   Foundations

   3 lessons

   1. 00aHow machines learn from examples
   2. 00bThe transformer insight
   3. 00cFrom architecture to LLM

   Explore module →
2. 

   Module 1モデル力学

   Model mechanics

   5 lessons

   1. 01From text to the next token
   2. 02Context is a finite workspace
   3. 03Long context is not perfect recall
   4. 04Embeddings: useful geometry, not truth
   5. 05Sampling: how one continuation is chosen

   Explore module →
3. 

   Module 2基盤システム

   Grounded and adapted systems

   3 lessons

   1. 06Hallucinations and reliability boundaries
   2. 07RAG: retrieve evidence before generation
   3. 08Choosing prompting, RAG, and fine-tuning

   Explore module →
4. 

   Module 3品質工学

   Quality engineering

   2 lessons

   1. 09Evals: define and measure useful behavior
   2. 10Testing probabilistic software

   Explore module →
5. 

   Module 4制御された自律

   Controlled autonomy

   2 lessons

   1. 11Prompt injection and secure tool use
   2. 12Agentic workflows with bounded autonomy

   Explore module →