Sentiment Classification

Built and deployed transformer models on 1M+ tweets with 95% accuracy for real-time sentiment analysis.

Diabetes Modeling

Developed logistic & random forest models achieving 85% accuracy for early diabetes detection.

Suicide Forecast

Created ARIMA & Prophet forecasting models with R²=0.85 for public health intervention planning.

Data Analyst @ Red Cross

Automated 500+ monthly receipt analyses using Azure PromptFlow, saving 10 hours/week and reducing manual effort by 85%, cutting costs by $10K/month.

---

Created Tableau dashboards for disaster response KPIs, boosting operational efficiency by 15% and aiding cross-functional decision-making.

Instructional Assistant @ UH

Conducted 30+ sessions for 50+ grad students on regression, hypothesis testing, Bayesian inference, and statistical computing with real-world data.

---

Led labs & provided personalized mentoring, resulting in 100% course completion and 85% scoring above 90%.