name: Capstone
number: 20
type: project
difficulty: advanced
estimated_hours: 15-20

description: |
  Final project combining all optimization techniques. Students build an optimized
  end-to-end ML system and compete on the global leaderboard.

learning_objectives:
  - Combine all optimization techniques
  - Build complete optimized systems
  - Deploy efficient ML models
  - Compete on performance metrics

prerequisites:
  - All previous modules (1-19)

skills_developed:
  - System integration
  - Holistic optimization
  - Production deployment
  - Performance engineering

final_projects:
  - Optimized CIFAR-10 trainer
  - Efficient GPT inference engine
  - Memory-constrained deployment
  - Custom optimization challenge