All Projects

Demand Forecasting System using Time Series AI

Data Science Project

Developed an end-to-end demand forecasting system to predict future sales using ARIMA and Prophet models.

  • Conducted time series analysis to identify trends, seasonality, and anomalies
  • Engineered lag features and rolling averages to improve model performance
  • Built and evaluated models using MAE and RMSE metrics
  • Visualized forecasts and insights using Python libraries
Python, ARIMA, Prophet, Time Series, Forecasting

AI-Powered IPL Auction Analysis & Price Prediction System

Data Science / Machine Learning Project

Built a machine learning system to analyze IPL auction data and predict player valuations from historical patterns.

  • Performed EDA to uncover player pricing trends and team spending patterns
  • Engineered features for predictive modeling
  • Built models to estimate player auction prices
  • Visualized insights using Matplotlib and Seaborn
Python, EDA, Machine Learning, Matplotlib, Seaborn

BeatBuff - AI-Based Music Recommendation System

Data Science / AI Project

Built a music recommendation system that analyzes user preferences and audio features to deliver personalized suggestions.

  • Processed music datasets and extracted relevant features
  • Implemented similarity-based recommendation algorithms
  • Improved recommendation accuracy through model tuning
Python, Recommendation Systems, AI, Data Science

Assent - Privacy-Focused Photo Sharing Web Application

Academic Project

Developed a privacy-focused web app enabling controlled photo sharing with identity-based access.

  • Designed secure authentication and access control system
  • Applied data-driven logic for identity-based content visibility
Web App, Authentication, Access Control, Privacy

Movie Recommendation System

Built a recommendation app using TMDB API and Streamlit with genre-aware suggestions and smooth browsing.

Movie Recommendation System

Built a recommendation app using TMDB API and Streamlit with genre-aware suggestions and smooth browsing.

Python, TMDB API, Streamlit, Data Science View on GitHub

Credit Risk Modelling System

Machine Learning / FinTech Project

Built a deployed credit risk platform that predicts borrower default probability, quantifies expected loss, and stress-tests entire loan portfolios.

  • Single Borrower engine computing PD & EL from financial inputs in real time
  • Portfolio stress-testing dashboard with PD multiplier simulation on a $41.6M book
  • Dynamic-programming FICO bucketing using log-likelihood optimisation
Python, Scikit-learn, Logistic Regression, Dynamic Programming, Plotly, Streamlit