kannan varshini

I've designed and implemented a real-time Scam Call Detection system using Content Analysis with Dynamic Sparsity Top-K Attention Regularization. I developed the end-to-end Java application, including signal processing, feature extraction, model integration, and alert generation.

Through my MCA final-year project at Annamacharya Institute of Technology and Sciences, I trained models on diverse call-content datasets to classify scam and legitimate calls. My work applies NLP, attention mechanisms, pattern recognition, and full-cycle software development to improve detection precision.

As an AI & Machine Learning Intern, I worked with real-world datasets, supervised learning, data preprocessing pipelines, predictive model development, and model evaluation. I also collaborated with a team to implement ML solutions and strengthen my professional communication.

I'm an MCA graduate with foundations in Python, Java, C++, SQL, data structures, algorithms, and object-oriented programming. I'm motivated to build intelligent, scalable solutions while continuing to grow with new technologies and frameworks.