Aditya Dhengare

American Express Credit Card Analysis

• Applied K-Means clustering on 500K+ credit card records to segment customers by spending behavior and identify high-value profiles; built classification models to detect fraud indicators with high precision using engineered features from transactional patterns. • Performed end-to-end data preprocessing in Python (Pandas, NumPy) — outlier treatment, feature scaling, and encoding — reducing data inconsistencies by ~25%. • Wrote complex SQL queries to extract and aggregate large volumes of transactional data, feeding clean datasets directly into ML pipelines. • Built 4 interactive Tableau dashboards to visualize model outputs, spending trends, and fraud risk scores, enabling data-driven decisions for business stakeholders.

ML Integration — Student Placement Prediction & Job Portal

• Designed and trained ML models (Logistic Regression, Random Forest) to predict student placement probability with 85%+ accuracy, using hyperparameter tuning and cross-validation; conducted EDA and feature engineering to identify key placement predictors. • Built a Flask backend to serve the trained ML model as a REST API, supporting real-time predictions for 200+ concurrent users. • Integrated Google Gemini API to enable AI-powered resume evaluation, generating personalized career recommendations for students. • Added an analytics dashboard to visualize placement trends, model confidence scores, and user activity, providing actionable insights for career counselors.