Justin Daludado

A LoRA model fine-tuned on the LeetCode Python dataset for solving LeetCode problems. Can train with either MLX or PyTorch.

A cleaned and merged dataset for supervised fine-tuning of LLMs on LeetCode Python solutions.

A Python code agent built with LangGraph that writes solutions for specific LeetCode problems, runs test cases, and self-corrects its solutions.

An MCP server that exposes Apple Developer Documentation in a structured format. Implemented it in Swift and distributed it through Homebrew and Swift Package Index.

A redesign of Carleton University's mobile app in SwiftUI with a cleaner structure and improved native UX. The project focuses on simplifying common student workflows and a smoother app experience.

ExploreRL is a SwiftUI app for iOS and macOS to visualize and experiment with reinforcement learning workflows. Integrated Gymnazo environments and MLX Swift to run and compare RL algorithms interactively.

A reinforcement learning toolkit written in Swift inspired by Gymnasium for Apple platforms. Added reusable environments, common RL algorithm implementations, and MLX integration for Apple Silicon acceleration.

A playback filter that blocks unwanted Spotify artists and tracks and switches to fallback sources. Implemented a headless Rust worker with a Next.js dashboard for configuration.

A FastAPI service that scrapes and serves Rakuten store and cashback data through a clean REST interface. Structured responses for simple integration with client apps.

An iOS travel companion app for discovering places and planning itineraries. Developed it in SwiftUI with Firebase auth and data storage, plus Google Maps and Places for location services.

Implemented post-quantum secure boot validation on Raspberry Pi 4B running QNX Neutrino RTOS. Integrated OpenSSL and wolfSSL with TPM-backed verification to harden boot integrity.

An autonomous parking agent with PPO in a custom Gymnasium environment. Used Python and PyTorch with curriculum learning and reward shaping to improve training stability and parking behavior.

A scheduling app for managers to create shifts and employees to view them in real time. Implemented role-based auth, availability-aware planning, and live updates using Next.js, React, Supabase, and PostgreSQL.

A full-stack social platform with authentication, posts, likes, follows, and image sharing. Developed the web app with Next.js, React, and MongoDB, and shipped a companion SwiftUI iOS app on the App Store.

A desktop grade manager to track courses, assignments, and performance. Developed it in C++ with Qt6 and SQLite to calculate course averages and GPA.

A real-time chat app where clients exchange messages over WebSockets. Implemented a Node.js and Express backend with a vanilla JavaScript client.

A Discord bot to automate Secret Santa signups, matching, and DM notifications. Implemented it with Python and discord.py, including duplicate prevention and dropout handling.