Pavanteja Reddy

I’m a detail-oriented Data Scientist focused on end-to-end machine learning pipeline development, predictive modeling, and data-driven analysis. I translate complex datasets into actionable insights using statistical analysis and strong data visualization.

I’ve built and evaluated classification, regression, and deep learning models with practical attention to feature engineering, preprocessing, and rigorous metrics. For a stock market forecasting project, I designed an LSTM-based time-series pipeline, engineered 10+ features, and improved RMSE by 15% versus a moving-average benchmark.

On the analytics side, I’ve worked with large, real-world datasets to uncover drivers and risk signals. In a loan default risk analysis, I analyzed 10,000+ records, identified 6 high-correlation features influencing default probability, and created dashboards for approval patterns, risk segments, and demographic distributions.

I also ship deployable ML solutions. I built a multi-class loan approval system achieving 88% accuracy with a tuned Random Forest, addressed class imbalance with SMOTE, and deployed the model as a REST API using Flask for real-time predictions.