Projects

Technologies: Python, Financial Modeling, Multi-linear Regression, Logistic Regression

• Reduced commodity storage costs by 8% through financial modeling with Python
• Boosted projection accuracy by 6% by applying multi-linear regression to forecast seasonal trends
• Engineered an advanced logistic regression model to evaluate credit risk
• Developed comprehensive data analysis pipelines for commodity market insights

Technologies: Next.js, Go, Gin, MySQL, Spotify API, JWT, OAuth 2.0

• Built a rhythm-based Spotify web application that generates playlists customized to running pace
• Implemented OAuth 2.0 authentication with Spotify and JWT session management
• Developed cadence-aware playlist generation using Spotify's recommendation API
• Created a modern UI with Next.js App Router, Tailwind CSS, and Radix UI

Technologies: Python, PyTorch, NumPy, CUDA, Transformers, LSTM

• Created NLP for sentiment analysis of 10 GB of X (Twitter) posts, leveraging CNN and transformer architectures achieving 92% accuracy
• Leveraged CUDA to dramatically accelerate the training of deep learning models, cutting training time by 100×
• Implemented custom data preprocessing pipelines, improving model accuracy by 20%

Technologies: C++, OpenGL, GLFW, GLM, Ray Tracing, Wave Physics

• Developed a real-time 3D wave simulation with realistic caustics rendering using OpenGL
• Implemented ray tracing algorithms for accurate light refraction and caustic pattern generation
• Created interactive wave disturbance system with physics-based wave propagation
• Built custom shaders for realistic water surface rendering and lighting effects

Technologies: Python, Selenium, React, Flask, Firebase

• Produced a full stack application using Python, Selenium, Flask, React, and Firebase
• Streamlined the collection of over 2,000 low-traffic Wikipedia sites using a Selenium web scraper
• Capability to scrape more than 3,000 unique sites per hour, significantly improving data acquisition efficiency