Hemanth R

EMR - Automated Medical Coding is an AI powered Medical Coding Automation Engine. Enhanced the accuracy of XGBoost models for EHR and EMR systems by integrating additional features. Mitigated data drift issues in models by optimizing the Random Forest algorithm and refining vectorization methods. Implemented object detection techniques to extract records from PDF files in the medical insurance domain. Utilized GPT-3.5 and GPT-4 to identify EM level specialty referrals and conducted experiments with LLMs, RAG, and Vector-DB.

New Car Demand Forecasting project. Developed Proof of Concept (POC) for new locations and channels. Gathered and pre-processed historical data for demand forecasting. Engineered features and selected relevant ones using PCA, Information Gain Ratio, and BorutaPy. Developed, evaluated, and optimized time series forecasting models including ARIMA, FBProphet, and UCM. Deployed models to AWS Sage-Maker Studio Endpoint and monitored model health.