Developed a Bidirectional LSTM model for predicting next day closing prices with a MAPE of 19% and built a dynamic portfolio optimization algorithm incorporating profit-taking and stop-loss strategies

This project builds a Music Genre Classification System using SVM, CNN, LSTM, and Transformer models, with YAMNet-based feature extraction. It processes audio files, extracts embeddings, and trains multiple models for comparison. The system predicts genres with high accuracy and provides evaluation metrics like classification reports and confusion.

📊 Time-Series Forecasting for Wearable Device Sensors

A full-stack hybrid stock prediction system using Bi-Directional LSTM. Integrates Technical Analysis, Macroeconomics (FED/Inflation), and Company Fundamentals into a single AI model with a Flask backend.

The project aims to predict MBTI personality types based on text data using deep learning models. It leverages LSTMs, Bi-Directional LSTMs, and BERT to enhance classification accuracy.