Swarna Reddy

Designed and deployed a scalable enterprise AI platform integrating Generative AI, ML, NLP, and big data engineering for investment research, risk analytics, regulatory compliance, and operational forecasting across asset servicing workflows. Built production-grade RAG systems using LangChain/LlamaIndex with Azure OpenAI and developed conversational copilots to enable natural language interaction with financial data and internal systems. Implemented end-to-end ML pipelines for risk classification, anomaly detection, and time-series forecasting using Python, PyTorch, and Scikit-learn, improving early risk detection and forecast accuracy. Engineered TB-scale big data pipelines using PySpark/Databricks/Spark/Delta Lake, and built FastAPI REST services to support AI applications. Implemented LLMOps/MLOps frameworks with Azure ML, Docker, and CI/CD (including monitoring with Azure Monitor/Application Insights), enabling reproducible deployments and drift detection. Applied Responsible AI an

Led design and deployment of a scalable enterprise AI platform integrating Generative AI, ML, NLP and big data engineering for investment research, risk analytics, regulatory compliance and operational forecasting. Built production-grade LangChain/LlamaIndex-based RAG pipelines and scalable APIs on Azure OpenAI, reducing research time by 40%. Developed end-to-end ML pipelines for risk classification, anomaly detection and forecasting, improving early risk detection accuracy by 27%. Implemented conversational AI copilots and integrated Bloomberg and Refinitiv APIs for real-time data workflows. Engineered high-performance big data pipelines with PySpark, Azure Databricks, Apache Spark and Delta Lake, processing TB-scale datasets. Built NLP/document intelligence models to extract entities, obligations and compliance risks from contracts and filings with 38% improvement in accuracy. Implemented MLOps and LLMOps using Azure ML, Git, Docker and CI/CD, achieving faster deployment cycles and a

Architected and scaled machine learning and generative AI solutions across Apollo Hospitals’ digital network to support clinical decision-making, disease risk prediction, and hospital resource optimization while maintaining HIPAA/NABH compliance. Built real-time ingestion pipelines using Apache Kafka and Airflow to synchronize clinical events (HL7 interfaces) and enable readmission forecasting. Used PySpark on GCP and BigQuery for large-scale processing. Developed supervised ML models (TensorFlow/PyTorch) for cardiovascular disease risk using feature engineering on FHIR-compliant EMR data. Built deep learning computer vision models (CNNs/RNNs, Keras/MONAI) for detecting pulmonary anomalies in medical imagery. Implemented LLM-based clinical NLP using Hugging Face Transformers for extracting diagnostic insights from physician notes. Applied evaluation and interpretability using scikit-learn metrics, XGBoost ensembles, SHAP/LIME (including R workflows) to reduce bias. Deployed inference