My Projects

A showcase of my work in Machine Learning, Signal Processing, and Full-Stack Development.

Nifty50 Stock Price Predictor

Built a 2-layer LSTM model for 10 Nifty stocks, achieving 51% directional accuracy. Engineered over 15 features including momentum, volatility, and correlation metrics. Deployed an interactive Streamlit dashboard with real-time prediction charts.

Python PyTorch Streamlit

Speech Segmentation & Word Discovery for Indian Languages

Developed a system for speech segmentation, word discovery, and phoneme recognition tailored to native Indian languages, addressing linguistic diversity and under-representation in technology.

Explainable AI NLP Speech Processing

Speech-Based Disease Predictor

Utilized the SpeechRecognition library to convert user speech into text for symptom extraction. Built a disease prediction model that compares extracted symptoms against a custom dataset of diseases using Pandas.

ASR NLP Transformers Pandas

Adaptive PCA-Based Portfolio Decomposition

Developed a quantitative pipeline analyzing global indices using Rolling PCA to extract latent factors and construct synthetic portfolios. Achieved a Sharpe ratio of 2.45 vs. the S&P's 0.63 with lower volatility.

PCA Quantitative Finance Portfolio Optimization