AB Experiment Design & Metrics

Rigorous experimental frameworks for product testing

About This Project

This project involves designing a rigorous experimental framework for product testing with multiple variants/features in a controlled environment. It includes calculating statistical significance for all test periods, assessing total earnings, and ensuring meaningful insights. The project also documents the full recommended pipeline for conducting experiments.

Core Concepts

  • A/B testing methodologies
  • Experimental design principles
  • Statistical significance calculation
  • Variance reduction techniques
  • Causal inference in experimentation
  • Documentation of experimental protocols

Key Knowledge/Skills

  • Experimental design
  • Statistical analysis
  • Hypothesis formulation and testing
  • Advanced randomization methods
  • Causal inference estimation
  • Reproducibility and best practices in ML

Coursework Covered

Statistical Methods for AI & Experimentation

Project Status

In development

Back to Projects