Shreyas Aundhekar

Built Generic Fraud score (XGBOOST & Random Forest) and Generic Microfinance scorecard (Logistic regression). Developed real-time recommendation engine for upsell opportunities. Created machine learning models using Python, pyspark, and sklearn to predict energy usage with 98% accuracy. Reduced third-party reporting costs by $100K+ per year.

Developed machine learning products for credit card customer strategy, covering churn prediction and campaign optimization. Provided risk consulting and business growth strategy for top banks in India and South East Asia. Built NLP model for customer feedback tagging and classified photos using deep learning (VGG16). Implemented demand forecasting models, improving accuracy by 45%.

Built rules using advanced machine learning techniques in Python. Developed 4+ use cases including lifetime value, early warning, and churn prediction. Automated quarterly value analysis using Pig Scripting and Python UDF, significantly reducing manual effort.

Model development using logistics regression and linear regression. Gained strong exposure to data preparation for model development and Random Forest implementation.

Model development using logistic regression, variable selection using Weight of evidence and Information value technique, and model validation using Gini and KS.

Worked on association analysis, clustering, and classifications.