KARINGA MADHUSHALINI

At Zydus-BC Solutions, I built a Python and Flask text-to-SQL prototype using Gemini 2.5 Flash for FDA pharmaceutical datasets, along with ETL scripts, SQL Server connectivity, validation checks, and reporting workflows. I also explored ML-based demand signals and VRP solver concepts for logistics planning across four hubs and 24 CFAs.

Previously at Edurva Learning, I developed Python and SQL machine learning workflows and improved model performance by approximately 12% through feature selection, preprocessing, validation, and metric-based evaluation. I'm currently researching AIOps-based anomaly detection for IT operations logs while building practical projects in data engineering, REST APIs, RAG, dashboards, and system monitoring.