Research

def merck(role="Undergraduate Data Science Researcher"):
    # built a semantic search engine over scraped, preprocessed PubMed chromatography abstracts (stopword removal, lemmatization)
    # combined SciBERT embedding cosine similarity with BM-25 into a hybrid relevance score, surfacing papers that shared no keywords with the query
    # designed the Neo4j knowledge graph behind it (7,000+ nodes, 11,300+ relationships), using GPT-3.5 to extract processes and materials from each abstract as graph entities
    # shipped it as a Streamlit interface returning ranked results with a percentage match per paper, validated on a 1,000-abstract corpus split evenly between related and unrelated documents
    # repo: github.com/tpaidich/ChromaSift
    ...

def nasa_space_biology(role="Undergraduate Data Science Researcher"):
    # built reproducible pipelines in R and SQL (MySQL) over NASA GeneLab multi-omics data
    # ran data-quality checks to quantify gene-expression shifts under varying gravitational conditions
    # documented methodology end-to-end for cross-team reproducibility
    ...