Session 1
Inference systems
Beginner
Understand and practice with no prerequisites
Session 2
How computers understand language
Session 3
The art of prompting
Session 4
AI can be wrong
Session 5
Multimodal AI
Session 6
AI in real life
Session 7
Rules of the game
Session 8
My first AI assistant
Session 9
AI and my data
Session 10
Final project
Session 11
How a model learns from examples
Intermediate
Build, evaluate, and deploy AI products
Session 1
Foundations: transformers and tokenization
Session 2
Professional prompting
Session 3
Structured output and evaluations
Session 4
RAG: model memory
Session 5
Tools and tool calling
Session 6
The agentic loop
Session 7
Multi-agent systems and MCP
Session 8
Infrastructure and deployment
Session 9
Building AI products
Session 10
Governance and final project
Session 11
The complete pre-training pipeline
Advanced
Production architecture, reliability, and optimization
Session 1
Claude API in depth
Session 2
Advanced tool use
Session 3
Agent SDK
Session 4
Multi-agent architecture
Session 5
MCP in depth
Session 6
Coding agents and CI/CD
Session 7
Advanced prompt engineering
Session 8
Context and reliability
Session 9
Architecture scenarios
Session 10
Practice exam and project
Session 11
Pre-training systems and optimization
Session 12
Q/K/V attention: from projections to causal output
Session 13
Linear attention and fixed-size matrix memory
Session 14
DeltaNet: correcting memory
Session 15
Chunking, causality, and parallel prefill
Session 16
Gated DeltaNet and selective forgetting
Session 17
Kimi Delta Attention: a bounded case study
Session 18
KV cache, recurrent memory, MLA, and low rank
Session 19
Mixture-of-Experts: routing and capacity
Session 20
Residual streams and attention over depth
Session 21