Identified, categorized, and documented software bugs by testing game builds using technical documentation, helping software engineers resolve critical issues before release

• Developed a responsive single-page portfolio website using pure HTML, CSS, and vanilla JavaScript to showcase personal projects, skills, and the proficiency needed to make the website without relying on libraries or frameworks.
• Integrated an animated background and dynamic contact form connected to Google Sheets for real-time messages while maintaining a clean UI and optimized performance across PC and mobile.

• Designing a full-stack e-commerce web app using JavaScript and the MERN stack in a team of 3, to allow users to register, purchase, and download 3D models via a secure database hosted on the cloud with MongoDB Atlas (AWS).
• Building RESTful API endpoints in Node.js and Express to handle secure HTTP requests, user authentication, and model management, while integrating rate limiting middleware and role-based access for admin-only controls.
• Developing a responsive frontend with React, Bootstrap, and modular JSX components for a clean user experience, while coordinating team progress using GitHub, Jira, and VS Code Live Share for real-time collaborative development.

• Built a two-layer neural network from scratch using Python in VS code with sigmoid activations, softmax output, and cross-entropy loss to perform digit classification.
• Implemented forward and backward propagation manually and trained the model via stochastic gradient descent, achieving over 94% test accuracy without using high-level machine learning libraries.
• Designed an interactive digit recognition GUI using Tkinter that preprocesses user input with cropping, antialiasing, and normalization to enable accurate real-time predictions from a trained neural network.