Evolution Log

System Evolution Log (Telemetry of Mastery)

When This Fires

• At the end of every Topic (mandatory)
• At milestone completion
• Manually via /evolution

The Evolution Log Format

Written to evolution/topic-N-evolution.md:

# System Evolution Log: Topic N → Topic N+1

**Date:** YYYY-MM-DD
**Sessions covered:** [session range]
**Time invested:** [hours]

## Code Complexity Comparison

### Topic N-1 Code
- Lines of code: [count]
- Number of classes/functions: [count]
- Test count: [count]
- Design patterns used: [list]
- Abstraction depth: [max nesting / composition depth]

### Topic N Code
- Lines of code: [count]
- Number of classes/functions: [count]
- Test count: [count]
- Design patterns used: [list]
- Abstraction depth: [max nesting / composition depth]

### Delta
- Complexity growth: [quantified]
- New patterns introduced: [list with first-use date]
- Patterns that became automatic: [list -- things the student does without thinking]

## Meta-Patterns Mastered

Identify 3-5 patterns the student has internalized across topics:

1. **[Pattern Name]** (e.g., "Dependency Inversion in AI Pipelines")
   - First encountered: [topic/session]
   - Became automatic: [topic/session]
   - Evidence: [specific code example where student applied it without prompting]

2. **[Pattern Name]**
   ...

## Agent Stress-Test Log

### Intentional Mistakes Used
Document any deliberate mistakes introduced to test the agent's coaching:
- What was the mistake?
- Did the agent catch it?
- How did the agent's recovery logic work?
- What does this reveal about the agent's reliability?

### Agent Coaching Quality
- Concepts where the agent excelled: [list]
- Concepts where the agent struggled: [list]
- Teaching approach changes made during this topic: [list]
- Model version used: [Claude model ID]

## Trajectory Evaluation

### Assessment Scores: Start vs End of Topic
| Dimension | Start | End | Delta | Evidence |
|-----------|-------|-----|-------|----------|
| Python Engineering | | | | |
| Problem Decomposition | | | | |
| Mathematical Thinking | | | | |
| ML/DL Understanding | | | | |
| System Design | | | | |
| Human-AI Collaboration | | | | |
| Engineering Craft | | | | |
| Evaluation Thinking | | | | |
| Communication | | | | |

### Velocity
- Exercises per session (average): [number]
- Time per exercise (average): [minutes]
- Trend: [accelerating / steady / decelerating]

### The "Click Moments"
List the 2-3 moments during this topic where understanding visibly shifted:
1. [What clicked, when, what triggered it]

### Open Gaps Carried Forward
- [Concept] -- deferred to Topic N+1 because [reason]

Why This Matters

At 18 months, the evolution/ directory contains a longitudinal record of growth that no interview can replicate. A recruiter reads topic-4-evolution.md and sees:

• Quantified complexity growth
• Named meta-patterns (not just "I learned stuff")
• Agent reliability data (the student EVALUATES the system that teaches them)
• Honest trajectory with click moments and open gaps

This is the "Telemetry of Mastery" -- the black box recorder of an engineer's development.