Dean Kuhn
I build data pipelines that end in a live, interactive product you can actually use.
Drug safety signal platform over FDA adverse event data (FAERS/openFDA). Cleaning and deduplicating two decades of shifting schemas, landing on Parquet + DuckDB, with a RAG layer over drug labels planned on top.
Bronze→Silver→Gold Delta Lake pipeline that correlated Reddit sentiment against Yahoo Finance price and volume data to flag S&P 500 divergence signals, surfaced in a Streamlit dashboard. Broken since Reddit shut down the public .json endpoints its ingestion relied on — kept here as a relic, not an active project.
CS graduate based in the Milwaukee area. I enjoy solving problems that involve cleaning and transforming difficult data, and displaying those insights in the form of live, queryable sites.
Each project runs the full stack: ingestion, transformation, warehousing, and, at times, machine learning.