PRAVEEN K

Integrated large healthcare datasets from EHR, claims, and patient monitoring systems to support operational and clinical analysis. Designed and optimized cloud-based ETL pipelines using AWS Glue, Redshift, Azure Data Factory, and Azure SQL Database, reducing processing time and improving accessibility. Developed ML models (Logistic Regression, Decision Trees) for patient risk prediction and used K-Means and PCA for patient segmentation. Optimized SQL queries and stored procedures in PostgreSQL and Redshift, cutting report generation time by 25%. Built interactive dashboards in Power BI and Tableau and leveraged Apache Spark (Scala/PySpark) on AWS EMR/Hadoop for HIPAA-compliant batch and streaming processing. Automated ETL and reporting workflows with Python and Excel VBA and implemented Snowflake tasks/streams for near real-time availability while reducing compute costs.

Collected and integrated IPEDS datasets, cleaning and transforming data with Python (Pandas), SQL, and Excel to prepare structured datasets for analysis. Optimized SQL queries and performed manual validation testing, reducing query runtime by 35% and improving analytics accuracy. Built dbt models and leveraged Hadoop (HDFS, MapReduce, YARN, Hive) to automate transformations, achieving 99% data accuracy and reducing ETL execution time by 30%. Applied forecasting techniques (ARIMA, Exponential Smoothing) in R and Python to improve KPI prediction accuracy by 15%. Designed a Text-to-SQL pipeline and RAG workflows using LangChain, PostgreSQL, and ChromaDB with OpenAI embeddings to deliver a low-code chatbot for non-technical users (95% accuracy), and delivered insights through Tableau dashboards and Lucidchart workflows.

Performed EDA using Python (Pandas, NumPy) and Tableau on global plastic waste datasets. Cleaned and preprocessed environmental data to maintain quality. Built predictive models with Logistic Regression and Decision Trees to forecast plastic waste accumulation by product and location. Evaluated models using precision, recall, F1-score, and confusion matrices, and created interactive Tableau dashboards visualizing waste trends by geography and brand. Applied statistical regression analysis and Pandas Profiling for automated EDA, then delivered recommendations to reduce plastic waste by identifying high-risk materials and regions. Documented methodologies to support environmental research.