My 2¢
2¢ is just too low. My opinions are worth slightly more in this inflation.
MIT CODE 2025: Validating Intuition & The Future of Experimentation
Validating intuition and the future of experimentation at the MIT Conference on Digital Experimentation.
The 'Litmus Test' Protocol: De-Risking Advanced Statistics
Why we validated CUPED in a Power Calculator before touching the Production Engine. A study in engineering precision.
The 'Sandbox' Strategy: How I Turned a Jupyter Notebook into an Enterprise App
Why you should scale your prototype before you build your platform. A story of turning a manual analyst loop into an automated engine.
From Sequential To Parallel: Scaling Experimentation Inference
A technical retrospective on re-architecting a data pipeline to reduce runtime by 75% and unlock $150M+ in revenue attribution.
Challenging Institutional Inertia: The $114M Friction Fix
Proved through XGBoost + Propensity Score Matching that employment verification had zero predictive value, unlocking $114M in annualized volume.
GenAI Experimentation Copilot
Eliminating 8+ hours/week of manual SQL toil by building an internal LLM agent that turns product questions into verified experiment analysis.
Interpretability IS Functionality: Reducing Authentication Friction by 13%
Why we chose Decision Trees over Deep Learning to stop account takeovers. A case study in surgical segmentation vs. black box models.