Data Scientist at Mastercard

Developed machine learning models for customer risk, transaction, and behavioral analytics using Python, SQL, and scikit-learn, improving prediction accuracy by 18% while reducing false positives by 22%. Designed end-to-end statistical analyses and predictive models on datasets exceeding 120M+ records. Built automated feature engineering and data preparation pipelines using SQL, pandas, and Python to clean, transform, and validate large datasets, reducing model development time by 35%. Evaluated and integrated LLMs and RAG to accelerate document intelligence workflows, reducing manual research effort by 45%. Developed AI-powered text analytics and NLP solutions to extract insights from unstructured customer feedback and operational documents, improving classification accuracy by 20%. Collaborated across 5+ cross-functional teams to translate analytical findings into actionable recommendations.

Data Scientist at Coforge

Developed and deployed regression, classification, and time-series forecasting models using Python and scikit-learn, improving forecast accuracy by 17% across client engagements. Performed EDA, feature engineering, and statistical hypothesis testing on structured and semi-structured datasets containing 30M+ records, uncovering key trends and improving model performance. Optimized complex SQL queries and analytical workflows to reduce execution time by 45% while improving data quality and reporting consistency. Designed and evaluated transformer-based NLP models for document classification, entity extraction, and information retrieval, reducing manual document processing effort by 35% and improving data accessibility. Applied SHAP-based explainability to interpret model predictions and strengthen stakeholder confidence.