• Overview

  • Training pipeline
  • Inference pipeline
  • Transformer evolution
• Foundations

  • Math cheatsheet
  • Core concepts
  • Training
  • Inference and KV cache
  • Components in depth
  • Attention block, shape by shape
• Systems

  • Parallelism
  • vLLM parallelism
  • Inference serving and vLLM V1
  • AI software stack