Selected work, ordered by release and active development.
A chronological archive of current and past projects across machine learning, product tooling, automation, web applications, and Minecraft Marketplace work.
Project Timeline
Recent systems first, older launches deeper in the archive.
Overview
This portfolio is a continuously evolving Next.js application designed to present engineering work as concise technical case studies rather than a static résumé.
Highlights
Built reusable project and content components for long-form engineering case studies.
Refined mobile lightbox behavior with dynamic viewport sizing, safe-area support, and responsive controls.
Designed the site around production work, measurable outcomes, and technical depth rather than generic portfolio sections.
Technologies
Next.jsReactTypeScriptCSSResponsive Design
Overview
This project was completed for IS 480: Advanced Business Analytics and explored whether engineered temporal network behavior could distinguish multiple classes of IoT attacks.
The dataset was derived from the Kitsune Network Attack dataset and contained more than 27 million packet-level records captured from a commercial IP-based surveillance and IoT environment under controlled attack scenarios.
Each packet was represented by 115 engineered behavioral traffic features extracted across five historical time windows. The analysis focused on supervised classification, class-level performance, and the usefulness of temporal behavioral features for attack detection.
Highlights
Processed and sampled from a 27M-packet IoT network dataset.
Built multi-class models for SYN DoS, ARP MitM, Mirai, and other attack categories.
Evaluated classification performance using precision, recall, F1-score, and ROC-AUC.
Analyzed feature importance across 115 engineered temporal traffic features.
OggSync is a desktop utility built to replace error-prone manual workflows involved in creating and maintaining Minecraft sound packs.
The application uses a state-driven PySide6 interface with asynchronous background services to scan asset trees, validate sound definitions, infer JSON mappings, surface errors quickly, and prepare assets for packaging.
The project emphasizes deterministic validation, fast feedback, and a workflow that remains manageable as sound libraries grow.
Highlights
Built a state-driven Qt interface with asynchronous background services.
Implemented recursive sound scanning, validation, and JSON inference workflows.
Reduced manual asset inspection by surfacing invalid definitions and missing files automatically.
Packaged the application as a standalone desktop executable.
Technologies
PythonPySide6QtJSONFile Systems
Overview
This project explored multimodal learning by pairing Minecraft gameplay imagery with human-written natural-language descriptions.
The work included dataset preparation, experimentation with convolutional and transformer-based approaches, and prototyping retrieval-oriented workflows in Google Colab.
The project is archived as an academic prototype rather than an actively maintained production system.
Highlights
Designed a data pipeline pairing gameplay samples with human-written captions.
Prototyped CNN and transformer-based model architectures in Google Colab.
Explored retrieval and evaluation strategies for caption-to-gameplay matching.
Technologies
PythonTensorFlowHugging FaceDeep Learning
Overview
NightPorter was built to reduce repetitive moderation work and improve onboarding consistency across Discord communities.
Features were organized into independent cogs with reusable configuration, logging, permission checks, and message templates so individual capabilities could evolve without tightly coupling the rest of the bot.
Highlights
Implemented role-based access control and reaction-driven role assignment.
Designed a modular cog-based architecture for feature isolation.
Added reusable message templates and structured operational logging.
Technologies
Pythondiscord.pyAutomationRBAC
Overview
UniMeet was a team-built university project focused on helping students form meaningful campus connections through structured onboarding and profile discovery.
The project is archived following completion of the course development cycle.
Highlights
Designed a multi-step onboarding and profile flow with safety-focused UX.
Implemented responsive, themeable UI components in Next.js.
Used Git-based team workflows and short development sprints to coordinate delivery.
Technologies
ReactNext.jsTypeScriptTailwind CSS
Overview
This application was developed as a clinical course project focused on accessible, low-friction recovery tracking for bariatric patients.
The system combined a patient dashboard, progress-oriented UI, shared PHP components, and straightforward forms designed around clarity rather than technical complexity.
Highlights
Built a patient dashboard with recovery and progress tracking widgets.
Created modular PHP components for shared page layouts and forms.
Collaborated in a small team using Git-based version control.
Technologies
HTMLCSSJavaScriptPHP
Overview
Emoji Pets is a social progression game centered on positive player interaction, customization, and lightweight progression.
Players earned likes through social interaction and used that progression to unlock buildings, emojis, furniture, and cosmetic options for their pets.
The project focused on approachable systems, expressive customization, and progression mechanics that rewarded interaction instead of competitive play.
Highlights
Built customization systems for emoji pets, including outfits and interactions.
Implemented progression mechanics driven by social engagement.
Collaborated with a team to design and ship a cohesive gameplay experience.
Technologies
Minecraft BedrockGame DevelopmentJSONJavaScript
Overview
Player Mobs explored an alternative approach to Minecraft entities by structuring mobs around player-like interaction and behavior patterns.
The add-on was developed under Nitric Concepts as a collaborative project and used custom entity definitions, behavior logic, and animation systems to create interactions that felt more expressive than standard mob implementations.
Highlights
Implemented custom mob behaviors using player-style entity structures.
Designed interaction systems that expanded how players engage with mobs.
Collaborated within a small team under Nitric Concepts to ship the add-on.
Technologies
Minecraft BedrockGame DevelopmentJSONJavaScript
Overview
One Block Skyblock reimagines the classic Skyblock formula around a single regenerating block that drives the complete survival loop.
Each block break advances a phase-aware progression system controlling block pools, loot, mobs, events, and difficulty while preserving compatibility across mobile, console, and PC versions of Minecraft Bedrock.
The project reached more than 1.7 million downloads on the official Minecraft Marketplace, earned a 4.8 / 5 average rating, and reached #1 most-played and most-downloaded map during its peak.
Highlights
Shipped a phase-driven survival experience built around a single regenerating block.
Reached #1 most-played and most-downloaded map on Minecraft Marketplace.
Reached 1.7M+ downloads with a 4.8 / 5 average rating.
Designed progression and content systems to perform across mobile, console, and PC hardware.