Adept at leveraging Python and AWS to drive data-driven decisions, my tenure as a Data Science Intern at Amazon empowered me to enhance Kindle and Alexa's user experience through sophisticated machine learning models. My analytical prowess and innovative approach resulted in significantly improved user engagement, showcasing my technical proficiency and problem-solving skills.
Project Summary
This project develops an online signature verification system using CNN and HOG for secure authentication Signature images are preprocessed, features are extracted using HOG, and classification is done via CNN Trained on 24 samples per signature, it employs KNN and Random Forest for prediction, determining whether a signature is genuine or forged The optimized model reduces false positives and negatives, evaluated using accuracy, precision, and recall Implemented in Python with TensorFlow and OpenCV, it enhances security in banking and legal verification