Created a tool that predicts famous celebrities from a scraped image dataset with 78% accuracy. Scraped over 1000 images from Google photos using python and selenium. Engineered features from the images of every Celebrity to identify face and eyes and cropped the perfect portion by using OpenCV(Haarcascades) and Wavelet. Optimized SVM, Random Forest, and Logistic Regression using GridsearchCV to reach the best model.

Created a tool using transfer learning that recognize fresh and diseased cotton leaves with 95% accuracy after running 20 epochs. Optimized Inception V3, and Resnet 50 to reach the best model. Built a client facing UI using JavaScript, HTML and CSS. Built a client facing API using flask. Uploaded this project on Cloud using Heroku. Cotton Disease Predicter UI This project’s GitHub Repository

Created a tool using Natural Language Processing that Analyze Stocks from a Kaggle dataset with 85% accuracy. Completed some EDA process and Join the Headlines together. Implement Bag of Words for Model Implementation. Optimized Random Forest to build the best model. This project’s GitHub Repository

Created a tool that predicts home prices from a Kaggle dataset with 85% accuracy. Engineered features from the dataset added new features also remove unwanted features using business logic and dimensionality reduction. Remove outlier using Standard Deviation and Mean also plot the data using Matlotlib to understand better . Optimized Linear Regression, Decision Tree, and Lasso Regression using GridsearchCV to reach the best model.