I design experiments, surface behavioral signals, and build the analytical systems that help product teams understand what is actually happening and what to do next.
Enterprise retail clients including Adidas, Zara, Dune London, and Crocs needed to know which recommendation strategies, pricing configurations, and campaign setups were actually moving revenue, not just engagement metrics. There was no structured experimentation program to answer that reliably.
A major retail client's recommendation click-through rate dropped noticeably over two weeks. The pipeline showed green. No errors flagged. The data looked structurally clean but the commercial signal was wrong and nobody upstream had noticed.
Commercial teams at Adidas, Zara, Dune London, and Crocs were pulling numbers from their own dashboards and arriving at different answers to the same questions. There was no single source of truth for engagement, conversion, or pipeline KPIs across the portfolio.
Program leaders were receiving delayed, manually exported reports. By the time data reached them, the window to intervene with at-risk participants had already closed. The analysis arrived after the problem had become permanent.
Built a production-grade streaming pipeline ingesting live Coinbase market data, processing it in real time using Kafka and Spark, and loading analytics-ready tables into AWS for downstream reporting and trend analysis. Demonstrates end-to-end ownership of a high-volume data system from ingestion through analytical output.
Non-technical stakeholders needed a way to query live data in plain English without opening a ticket or learning SQL. The challenge was making it safe: no write access, no schema hallucinations, no unvalidated results reaching the user.
Excited about building analytics that helps creative entrepreneurs grow their businesses.