Nandini Katta

Currently working on enterprise Generative AI solutions for financial document intelligence and knowledge search. Built RAG pipelines, vector search, and document processing workflows using Azure OpenAI, Azure Cognitive Search, Python, Databricks, and PySpark. Focused on improving retrieval quality, grounded responses, summarization, and reliable AI outputs for business users.

At Optum, I designed and implemented AI/ML solutions for healthcare analytics, including fraud detection, patient risk stratification, and claims optimization. I built NLP pipelines using transformer models like BERT and ClinicalBERT to process clinical notes and extract meaningful insights for decision support. I also worked on end-to-end MLOps workflows using AWS and Azure, and developed scalable data pipelines with Snowflake and AWS Glue. Along with this, I contributed to building predictive models and real-time AI systems to improve healthcare outcomes and operational efficiency.

At MetLife, I built and deployed machine learning models for customer churn prediction, policy lapse forecasting, and fraud detection to improve underwriting and risk decision-making. I worked on segmentation models and predictive analytics to identify high-value customers and optimize business outcomes. I also developed time-series forecasting models for claims and renewals, and applied NLP techniques for analyzing customer feedback and agent notes. Along with this, I contributed to automated data pipelines and model evaluation frameworks to improve performance and reliability in production systems.

At Microsoft, I developed and deployed machine learning models for customer analytics, forecasting, and NLP use cases like churn prediction, behavior analysis, and demand forecasting using XGBoost, Random Forest, and LightGBM. I also built NLP pipelines for text classification, semantic search, and document intelligence using BERT and transformer models, along with end-to-end ML workflows covering data processing, feature engineering, training, and deployment.

At Edvensoft Solutions India Pvt. Ltd, I developed and deployed machine learning models for credit risk scoring, churn prediction, and loan default detection. I worked on exploratory data analysis, feature engineering, and building classification models using Logistic Regression, Random Forest, and XGBoost to improve decision-making accuracy. I also designed recommendation systems, implemented A/B testing frameworks, and built explainable AI solutions using SHAP and LIME. Along with this, I worked on cloud-based ML deployments using AWS and Azure, and created automated pipelines for data preprocessing and model training.