At NEWME, I built a D7/D14/D20 lifecycle tracker in SQL and Metabase that flags the bottom 10% of launches by D14. It replaced a process with no early performance check, helping merchandising make restock, markdown, and retire decisions sooner as sell-through rose from 40% to 48%.
At NEWME, I also broke down order mix, AOV, and item-level revenue by category for BNM, exposing a 12% AOV gap that shifted merchandising priorities. My age and ASP segmentation found that customers aged 25 to 34 spend 18% more per order, informing their pricing and promo plan.
At Tata CLiQ, I built an 80:10:10 CRM holdout test that showed reported lift was overstated by 30%, leading to corrected GMV reporting. I also moved three brand pipelines from S3 to Redshift with Python and Boto3, cutting ingestion time by 40% and enabling same-day Power BI reports.
My projects include a Python fraud detection model and Power BI dashboard, as well as a customer churn risk analysis using Python and SQL across 64K+ records. I work across analytics, experimentation, and data pipelines to help teams make decisions from their data.

