Madhumitha Kuppachi

Women in Data Science Accelerator Program

Accenture•  March 2024 - April 2024•  Dublin, Ireland

Collaborated with a team of 4 students to build a predictive model for fraud detection in Insurance Claims, analyzing over 20000 data points to identify fraudulent activities. Employed machine learning techniques such as logistic regression, decision trees, and random forests to achieve a classification accuracy of 92%, with precision of 85% and recall of 88%. Created data visualizations using Power BI to communicate findings to stakeholders, reducing reporting time by 20%. Identified 10 features leads to Fraud Detection and recommended investigations, leading to an estimated cost savings of €63,423,738. Presented findings and recommendations for cost reduction to a simulated executive board (acting CTO, CEO, and CFO), emphasizing the business impact of improved fraud detection.

Data Scientist Intern

Orcawise•  October 2023 - March 2024•  Dublin, Ireland

Designed a BERT model for Entity Extraction and finding Multi Class relationships, achieving an outstanding 85% overall accuracy on diverse news data types, showcasing proficiency in model development and successful implementation. Investigated on Health Insurance Portability and Accountability Act Dataset and reduced the dataset using statistical techniques by 20% and fine-tuned using Open AI GPT 3.5 Turbo for detecting disease severity.