Pranathi Deepak

Built batch and near real-time data pipelines using Python and SQL to prepare model-ready datasets, reducing manual processing effort by 60%. Developed data validation and monitoring workflows using Pydantic and PyTest to reduce downstream data errors by 30%. Created reproducible feature engineering processes to support consistent model training and inference. Supported end-to-end ML lifecycle workflows including data ingestion, preprocessing, and deployment preparation using AWS services. Maintained version control and CI/CD processes using Git to improve reliability and deployment consistency.

Taught algebra and basic statistics concepts to 15+ students, improving average exam scores by 20%. Developed structured problem-solving exercises focused on equations and probability concepts to strengthen analytical thinking. Used data tracking methods to monitor student progress and adjust lesson plans based on performance trends.

Built scalable ML pipelines using Python, Spark, and AWS to support batch and near real-time inference workloads. Developed supervised and unsupervised models using scikit-learn and XGBoost with cross-validation and hyperparameter tuning. Designed feature stores in Redshift and improved Spark and Databricks processing efficiency, increasing SLA compliance by 25%. Implemented CI/CD workflows and automated testing using PyTest and Git to improve deployment stability. Evaluated model performance using ROC-AUC, precision, and recall metrics to ensure reliable production deployment.