About

Most user analytics are reactive, but ZervePulse is proactive. I built a LightGBM-powered classification pipeline that analyzes a user's behavior within their first 24 hours to predict long-term platform success with 99% precision. By engineering features focused on "AI Agent Interaction Velocity" rather than just session time, I identified that iterative tool usage is the "North Star" metric for retention. The final result is a production-ready API that allows Zerve to automate onboarding for high-potential users before they ever hit a friction point.