Pinned Loading

1. Heart-Disease-ML-Pipeline Heart-Disease-ML-Pipeline Public

   Gradient Boosting heart-disease risk classifier reaching 0.954 AUC and 86.6% sensitivity across 630,000 clinical records (13 features).

   Python

2. ML-Smoker-Status-Prediction ML-Smoker-Status-Prediction Public

   Binary classifier predicting smoking status from clinical biomarkers using Logistic Regression, Random Forest, and Gradient Boosting (best AUC: 0.8887).

   Python

3. Drug-Interaction-Severity-Prediction-Multi-Task-ML-Cluster-Analysis Drug-Interaction-Severity-Prediction-Multi-Task-ML-Cluster-Analysis Public

   Predicts drug-interaction severity and adverse-event risk from FDA safety data. Multi-task model reaches 0.925 AUC across 15,000 records using 126 engineered features.

   Python

4. map-elt-research-graph-3 map-elt-research-graph-3 Public

   Concurrent Python pipeline profiling 22 companies in parallel across SEC EDGAR, ClinicalTrials.gov, NIH RePORTER, PubMed, and OpenAlex with zero LLM calls.

   TypeScript 1

5. prescription-event-stream prescription-event-stream Public

   Checks every prescription against federal dispensing rules and screens for drug interactions. Same finding, three views: patient, pharmacist, and prescriber.

   Python