Vinay Kumar

Experienced in model risk management, with expertise in writing model codes, conducting data sense checks, and performing comprehensive analyses to ensure accurate and robust models for validating financial models throughout their lifecycle. Proficient in Python, R, Excel, and Word for supporting model validation tasks. Strong background in ensuring compliance with internal policies and external regulatory requirements. Capable of producing detailed validation reports to effectively communicate findings and recommendations.

Airport Revenue Management (IIT DELHI Thesis): Used Non-Aeronautical Data from Delhi International Airport Limited (DIAL) and performed Exploratory Data Analysis (EDA), and applied Data Analytics, Machine Learning techniques and Market Basket analysis to predict the spending band for customers along with the category of product on which customer will spend the most. Finally deriving the conclusions from the analysis done into data-driven actionable commercial insights and recommending marketing techniques to promote sales and revenue. 

Customer Churn Prediction using ML Algorithms: Dataset contained 3333 entries of customers with 19 features, it is imbalanced with 86% non-churn entries. Performed EDA, Label Encoding, Feature Scaling and Data visualization techniques on the data. Applied SMOTE balancing & Logistic Regression, SVM, KNN, Random Forest, XGBoost ML models. Concluded XGBoost with 95.8% accuracy, 93.0% precision, and 89.0% recall is best model for given data. 

Stock VWAP Prediction using Time Series: Predicted two years of VWAP from the past 20 years of Reliance stocks data using Time Series Forecasting. Decomposed time series to analyze Trend & Seasonality, Plotted ACF and PACF for determining p and q. Tested Stationarity using Augmented Dickey-Fuller test, applied Differencing to make the data stationary. Applied MA, ARIMA, SARIMA models on the given time series data to predict VWAP. Concluded that SARIMA model gave better results i.e. 102.14 RMSE & 0.06% MAPE.