Design, develop, and deploy AI/ML solutions for healthcare analytics, clinical operations, claims intelligence, patient engagement, and operational decision support. Built NLP pipelines to extract structured insights from unstructured clinical notes and documents using Python, Hugging Face Transformers, spaCy, and SQL. Developed ML models for patient risk scoring, claim anomaly detection, readmission prediction, and care-gap identification. Implemented an end-to-end Retrieval-Augmented Generation system using LangChain, OpenAI/Azure OpenAI APIs, embeddings, and FAISS/Chroma for semantic search with grounded answers and citations. Built embedding ingestion pipelines for PDFs and operational documents, improving retrieval speed and reducing manual review time. Created training/evaluation pipelines with Scikit-learn, PyTorch, MLflow, and Airflow, plus FastAPI inference services and dashboards. Implemented monitoring workflows for drift, data quality, latency, and performance degradation,

Machine Learning Engineer at ICICI

August 1, 2020 - July 1, 2022

Contributed to banking analytics, fraud detection, credit risk modeling, customer segmentation, and operational automation. Built supervised models for fraud detection and suspicious activity classification using Python, SQL, Scikit-learn, XGBoost, Random Forest, and logistic regression. Engineered features from large-scale financial datasets using SQL, Pandas, and PySpark. Developed credit risk and customer behavior models and anomaly detection for unusual transaction/account patterns. Built customer segmentation using K-Means and DBSCAN. Produced evaluation reports (confusion matrix, ROC-AUC, precision/recall/F1, lift charts) and automated recurring risk/compliance reporting with Python and SQL. Developed Power BI/Tableau dashboards to monitor fraud trends and portfolio risk. Supported deployment workflows and post-deployment monitoring/retraining analysis, improving fraud precision, reducing false positives, and accelerating risk review prioritization.