I build end‑to‑end analytics systems, AI tools, and polished dashboards – from raw data and video to insights people can actually use.
Recent focus: sports analytics, AI agents, and interactive BI dashboards.
Shipping complete solutions – not just models or reports. I like projects where I can own the data, the code, and the final experience people see.
A mix of engineering, analytics, and design so projects don’t stop at a notebook – they become tools people actually use.
Six advanced AI & ML projects — deployed pipelines, computer vision, GenAI, and sports analytics.
Real-time in-game win probability with SHAP explainability. XGBoost + isotonic calibration trained on 500K+ plays. Deployed FastAPI endpoint + live Streamlit dashboard.
Predicts 14-day soft-tissue injury risk from workload patterns. Cox Proportional Hazards model for time-to-injury + LightGBM classifier. SMOTE handles class imbalance.
Natural language Q&A over 10 years of NBA player stats. Embeds player data into ChromaDB, retrieves via cosine similarity, and feeds context to Claude/GPT-4o for grounded answers.
Production-grade ML pipeline with drift detection, auto-retraining, and Grafana monitoring. Daily Prefect flow triggers retrain when PSI > 0.2. Full Docker Compose stack.
Upload basketball footage and get an annotated video + tactical JSON report. YOLOv8 detects players, ByteTrack assigns persistent IDs, court homography maps positions to 2D.
Predicts career WAR, All-Star probability, and bust risk from college stats and combine data. Includes an interactive trade analyzer with ML-derived pick value curve.
All public repositories pulled live from GitHub — ML models, analytics pipelines, dashboards, and experiments.
Want to talk about a role, a project, or collaborating on analytics work?