A desktop scheduling app built in Java Swing, designed around goals and progress rather than events on a calendar. Users break a Module (a time period) into Sub-modules (goals), Tasks, and Subtasks, marking what they complete. A separate Python pipeline reads that activity on a schedule and computes summaries — completion trends, streaks, and goal progress — which feed an Insights screen showing how the user is doing over time. Both the live operational tables and the analytics summaries are stored in PostgreSQL. That data-engineering layer is the core of the project and the thing a standard calendar app doesn't do.

GitHub

A 64-bit turned based fantasy video game about exploring dungeons, defeating monsters, and gaining loot. Heavily inspired by Dragon Quest. The game is story driven, contains stats, contains custom moves, loot dropping, dodging mechanics, and challenging bosses.

A full-stack, web-based point-of-sale system for a bubble tea shop, unifying every touchpoint of restaurant operations: a customer self-service kiosk, cashier POS, kitchen order queue, dynamic menu board, and a manager dashboard with sales analytics and inventory tools. Built with React, Next.js, and PostgreSQL, the system features role-based access for managers, cashiers, and customers, plus accessibility-focused design following WCAG guidelines.

As project manager and backend developer on a six-person team, I coordinated Agile sprints, authored SQL queries for order processing and reporting, and built checkout, cart, and kitchen UI components.

GitHub

A progression through core computer graphics techniques, from writing a software rasterizer by hand to building a full ray tracer. Assignments cover both CPU-side math (rasterization, ray casting) and GPU-side programmable pipelines (GLSL shaders, deferred rendering), with increasing complexity in geometry, lighting, and animation.

• A1 — Software rasterizer: triangles, z-buffering, per-vertex and interpolated color
• A2 — Hierarchical scene graph: joint-based character with parent-child transforms
• A3 — GLSL shading: Blinn-Phong lighting, silhouette edges, and cel shading
• A4 — 3D scene: freelook camera, texture mapping, and animated objects
• A5 — Deferred rendering: G-buffer storing position/normal/color for multi-light shading
• A6 — Ray tracer: shadows, mirror reflections, mesh intersection, and motion blur

A small personal project that allows the user to test their typing speed. This project contains multiple modes, this includes: prompts with random words, prompts with a scatter of random letters, and improving prompt mode (basically keeps track of letters/symbols that the user frequently misses and makes a new AI generated prompt based on that). Heavily inspired by Monkey Type.

GitHub

Planned — AI Scheduling Assistant. Layered on top of the completed scheduler, this upcoming feature turns the app's own usage analytics into personalized planning help. It reviews your tracked behavior — completion rates, how often tasks get pushed, when you actually get work done, and which subjects lag — and recommends concrete schedule adjustments. It also lets you manage tasks conversationally ("put this on next Tuesday"), asking follow-up questions when it needs more detail before acting. Because it reasons over your real history and every action it proposes is validated by the app before touching your data, the assistant stays grounded rather than guessing. (Planned feature, not yet implemented.)