Vaishnav Kumar Chandha

Designed credit-scoring and delinquency-prediction models (Logistic Regression, XGBoost) with Risk, Quant, and Engineering teams, improving risk-forecast accuracy by 18% and strengthening portfolio resilience and loss-mitigation strategy. Built and deployed fraud-detection and anomaly-monitoring pipelines, reducing false positives by 22% and enhancing real-time transaction surveillance across institutional and retail portfolios. Engineered financial KPI forecasting models for capital-adequacy and liquidity planning, improving reporting precision by 25% and supporting strategic balance-sheet management. Architected and automated AWS data pipelines (Glue, Redshift, Lambda) processing 2M+ daily records, cutting data latency by 40% and ensuring timely, accurate ingestion for model training and reporting. Implemented Generative AI and LLM-powered RAG pipelines to automate credit-policy summarization, compliance alerting, and risk-report generation, decreasing manual review time by 35%. Part

Developed patient-demand and resource-utilization forecasting models (Prophet, ARIMA, LSTM), improving admission/discharge forecast accuracy by 25% (RMSE) and reducing bed and staff over-allocation costs by 15% across departments. Built patient-churn and risk-prediction models (XGBoost, Random Forest, K-Means), increasing patient retention by 30%, improving follow-up adherence by 20%, and enabling personalized care-outreach programs. Led A/B testing and statistical experimentation with clinical, telemedicine, and marketing teams to evaluate dynamic appointment pricing and outreach campaigns, raising patient conversion and engagement by 6%. Engineered real-time anomaly-detection pipelines (Isolation Forest, streaming analytics) to flag abnormal patient-inflow patterns, reducing ER bottlenecks and minimizing service downtime. Deployed scalable Azure data pipelines integrating EHR, patient-feedback, and billing data for unified analytics and HIPAA-compliant AI model deployment. Built inte