[MR Train 1/5] A/B Testing: Configuration framework

MR Train: A/B Testing Framework for Recommendations

MR 1 of 5 - Building experiment infrastructure

Train Structure

1. Configuration framework ← YOU ARE HERE
2. Variant assignment engine
3. Metrics collection
4. Analysis and reporting
5. API integration

This MR

Adds the configuration layer for defining and managing A/B test experiments.

Changes

• ExperimentConfig and Variant models
• ExperimentRegistry for managing experiments
• YAML configuration support
• Traffic allocation validation

Design Decisions

• Using YAML for easy experiment definition
• Strict validation (traffic must sum to 1.0)
• Exactly one control variant required

Related

Part of #20 - Improve recommendation CTR by 25%

@jean_gabriel @sabrina_farmer Please review the config design!