For a decade, I've built the analytics layer that tells teams whether their decisions are actually working — rigorous A/B tests, complex SQL, and Media Mix Models that quantify what's really driving business outcomes. Now I'm building the engineering layer underneath: production AI systems, from RAG pipelines to fine-tuned models, deployed end to end.
I'm a Manager of Advanced Analytics at WPP Media, where I build MMM and attribution work on Unilever-scale brands, and before that a Senior Data Scientist at Coinbase. In parallel I've built eight AI agents that cover RAG, MCP, LangGraph, CrewAI, multimodal input and LLM fine-tuning, and I ship them the way a production team would: AWS Lambda, Terraform, CI/CD, tracing with Langfuse, and experiment tracking with MLflow. Below you'll find the code for each one, and on the right you can ask my AI twin anything about my background.
Agent portfolio
Production
How I ship, trace and measure LLM systems: the infrastructure, the tracking and the evaluation around the model.
About
10 years across marketing analytics, data science and data engineering. Currently Manager, Advanced Analytics at WPP Media (formerly GroupM / Mindshare) in New York, building MMM and attribution work on Unilever-scale brands. Before that, Senior Data Scientist at Coinbase.
M.S. Business Analytics, Robert H. Smith School of Business, University of Maryland. B.Tech in Mechanical Engineering. My core strengths are experimentation (A/B tests and hypothesis testing), SQL at scale, media mix modeling, and turning models into dashboards and narratives that executives act on.
I'm building this portfolio independently, in parallel with my day job, toward ML Engineer / Applied AI Engineer and MLOps roles, with an edge toward MarTech and AdTech. That means pairing the analytics side with hands-on engineering: RAG, agents, fine-tuning, Docker, Terraform, Kubernetes, CI/CD and observability.
Download resume →