KaggleProjects

• Titanic: Machine Learning from Disaster competition - Developed model which ranked Top 6% Where in employed feature engineering to better expose data and parameter tuning.
• Digit Recognizer - Implemented basic neural network on MNIST data with 97% accuracy.
• Microsoft Malware Prediction – Employed Dask to deal with huge dataset (8 million rows) using parallel processing and various imputations to handle missing values. Used Xgboost, ensembled decision trees algorithms.
• House Prices: Advanced Regression Techniques - Regression model predicting the price of home based on the features and facilities.