Projects

A project used for securely embedding and extracting hidden messages inside image files. It implements salted hashing (SHA-based) to verify message integrity, applies pixel-level steganography for covert data storage, and includes a lightweight GUI for practical use. Designed for simplicity, security, and consistent extraction without quality issues.

A project for scalable stock-market sentiment analysis that ingests news and market data, computes sentiment-driven trading signals, and deploys them through a containerized pipeline using Docker and Kubernetes for distributed processing and reliability.

A project that hosts Team Stonkholm’s complete submission for IMC Prosperity 3 across all five rounds - featuring statistical arbitrage, market-making, and hybrid trading strategies, along with our full trading engine, evaluation utilities, and final delivery scripts.

A project to maintain a modular, version-controlled checklist for tracking Formula Student rules (uses Formula Bharat 2024 CV Rulebook) and ensuring subsystem-wise compliance.

A project implementing a pairs-trading strategy that identifies cointegrated asset pairs, models their spread behavior, and generates market-neutral long/short signals. Includes backtesting, performance evaluation, and parameter tuning to assess profitability and risk across varying market conditions.

A project that implements the Heston stochastic volatility model to simulate asset prices and variance over time. The project explores the dynamics of asset prices under varying volatility structures and compares the results with traditional models like Geometric Brownian Motion (GBM).

A project focused on modeling market risk using Value at Risk (VaR) and Conditional Value at Risk (CVaR) to assess potential losses in a portfolio. The project uses historical simulation and parametric methods to calculate risk metrics and visualize tail risk exposure.

A project dedicated to constructing the yield curve for U.S. interest rates using over the years. By applying the Nelson-Siegel and Nelson-Siegel-Svensson modeling techniques to analyze interest rate trends and understand the dynamics of the U.S. market.

A project focused on extracting NVIDIA stock data from Yahoo Finance and developing new features to enhance model building. By combining relevant data points, the project aims to uncover insights into the stock’s performance.

A project that uses linear regression to predict the Nifty50 index based on historical data, involving some exploratory data analysis (EDA) to see how well the model fits the price movements.

This project explores advanced seismic data visualization techniques using Python. It leverages libraries like Mayavi and Matplotlib for interactive 3D rendering, attribute mapping, well log integration, structural interpretation, geobody modeling, seismic inversion, AVO/AVA analysis, time-lapse seismic, and geostatistical analysis.

This project builds a real-time algorithmic trading system using Apache Flink, Apache Kafka, Python, SQL, and Docker, integrating Alpaca API for market data and trade execution. It features real-time data processing, custom trading algorithms, and sentiment analysis.

This project aims to analyze sentiment in Amazon Alexa product reviews by classifying them as positive or negative. Using machine learning models like Random Forest, XGBoost, and Decision Trees, the project preprocesses text data, extracts features, and evaluates model performance through accuracy, confusion matrices, and cross-validation.

This project analyzes and compares machine learning models using FP32 and BF16 precision. It includes data preprocessing, model training, evaluation, and performance analysis, demonstrating the computational benefits of BF16 without sacrificing accuracy.

This project focuses on building a Learning Management System (LMS) using NextJS and ShadCN, designed to streamline the management of online education. The system includes user-friendly interfaces for both students and instructors, enabling seamless course creation, enrollment, and tracking.

This project utilizes machine learning, specifically a Random Forest Regressor, to predict the Factor of Safety (FOS) for materials used in vehicle chassis design. By analyzing material properties such as tensile strength, modulus of elasticity, and density, the model predicts FOS to help select the most suitable material for design engineering.