Developed an end-to-end clinical survival prediction system for LVAD patients, combining deep survival modeling, robust preprocessing, and explainable AI for decision support Built and interpreted time-to-event models (DeepHit) and generated clinically grounded explanations using SHAP/LIME-style XAI analysis Created an interactive Streamlit-based decision support application with: Natural language query processing via LLM integration, individualized risk profiling, survival-curve visualization, and explanation-driven scenario analysis