PRABHATH VINAY VIPPARTHI

Worked with business and AI requirements, building Python/SQL/PySpark pipelines to process 10K+ records and reduce manual validation effort. Prepared training and evaluation datasets and engineered statistical/ML features for anomaly and data-quality detection. Built classification models using scikit-learn and PyTorch with cross-validation, precision/recall analysis, and threshold tuning to improve minority-class recall. Integrated Hugging Face transformers and embeddings for text classification and semantic evaluation, implemented embedding-based retrieval workflows, and exposed REST APIs for similarity search and LLM evaluation. Supported deployment using MLflow, Docker, Kubernetes, Azure, and Databricks, and monitored quality, drift, latency, and pipeline failures. Built Power BI dashboards for AI/ML performance, data-quality trends, and business KPIs.

Translated retention and growth objectives into churn and forecasting analytics use cases, including scoping analytical approaches. Performed EDA and statistical analysis using Python (pandas/numpy) and SQL to uncover customer behavior patterns and churn drivers. Engineered 30+ customer-level features (e.g., recency, frequency, transaction value, engagement) to improve predictive modeling inputs. Built and compared Logistic Regression, Random Forest, and Gradient Boosting classifiers using cross-validation and ROC-AUC, improving high-risk customer recall. Developed demand forecasting models from historical and seasonal patterns to reduce forecast error. Processed 5K+ transaction records with PySpark/Databricks, automated recurring workflows with Airflow/AWS/MLflow, and delivered Tableau dashboards covering churn segments, forecasts, and business KPIs.