Key Projects & Technical Work

ShapeSite Studio is a macOS app for building complete websites visually, then exporting real deployment-ready code.

Key capabilities:

• Drag-and-drop visual editor with advanced canvas controls (grid, snap, guides, spacing/debug overlays, zoom, multi-select, alignment, grouping, layers).
• Flexible layout system: freeform, auto layout stacks, grid, constraints, and responsive breakpoints.
• Rich design tooling: geometric + organic shapes, vector path editing, typography controls, styling/effects, and asset management.
• Built-in component system: reusable blocks, custom components, variants, and plugin-like components with schema-driven settings.
• Live preview + inspect workflow: select/edit elements directly from preview and keep visual/code workflows in sync.
• Internal code editor for HTML/CSS/JS with split-mode editing.
• Import existing HTML/pages (rule-based DOM/CSS parsing, no AI required) and convert to editable structures.
• Data/API binding and dynamic runtime scaffolding with secure proxy templates.
• Optional admin/auth/database scaffolding for static + backend-ready exports.
• Security toolkit with hardened presets, security audits, host-specific config generation (Netlify/Vercel/Nginx/Apache/etc.), and deployment guidance.
• Project system with autosave, version timeline, checkpoints, and portable .shapesite files.

In short: it combines visual design power (Figma/Webflow style) with practical engineering output (clean HTML/CSS/JS + deploy/security scaffolding) in one local macOS studio.

Designed and developed a multi-purpose autonomous drone platform intended for emergency response, environmental monitoring, and autonomous delivery applications. The system integrates onboard GPS, LiDAR, barometric pressure, temperature sensors, cameras, and audio communication for real-time data collection and remote situational awareness.

The drone includes a mechanical gripper system capable of transporting and deploying payloads such as medical supplies, food, or emergency equipment to hazardous or difficult-to-access locations.

Autonomous flight capabilities were implemented using ArduPilot and Mission Planner, including preprogrammed GPS waypoints, obstacle avoidance, and automated landing for reliable low-intervention operation.

A Raspberry Pi companion computer supports onboard processing, sensor coordination, and data collection. The drone functions as a mobile sensing platform for environmental mapping, aerial imaging, atmospheric monitoring, and two-way communication.

(Honors Thesis Project, includes both hardware and software creation.)

Developed a multi-component intrusion detection system featuring a Java-based security analytics dashboard with a prototype Next.js web interface for future web-based monitoring.

The platform aggregates and correlates security events from system resources, network traffic, file-system activity, and threat intelligence analysis modules for real-time visibility.

System Monitor tracks CPU, memory, and processes via OSHI telemetry. Network Monitor captures/analyzes traffic to detect unusual behavior. File Monitor tracks file changes with suspicious-pattern scoring and quarantine support.

Dashboard was built with Java Swing and JFreeChart, with real-time tabs/charts/event logs. An event-driven architecture using BlockingQueue allows modules to feed a centralized analytics engine with filtering/search/prioritization/historical analysis.

AURA is an all-in-one productivity and intelligence app that helps users:

• Sync assignments and manage deadlines
• Prioritize tasks and generate AI-based schedules
• Chat with AI for study/work support
• Organize documents, recordings, and transcripts into sections/subsections
• Ask AI questions grounded in uploaded content
• Get field-specific security/videos/news intel/events
• Discover jobs and learning resources
• Track productivity trends and wellness signals
• Use a Chrome extension to capture web insights into the app

GuardiaPass combines encrypted credential management, browser protection, and intelligent assistants in one system.

It secures passwords, identities, and sensitive media with local encryption, then extends protection into the browser with autofill, phishing/domain checks, breach intelligence, and suspicious-email filtering.

Includes multiple AI assistants: in-app Security Advisor, extension page assistants for live site analysis, and contextual tools for password generation/strength/risk explanations.

Accessibility includes read-aloud support. Architecture includes local-first secure storage, server-side API proxying for key protection, real-time telemetry, trust checks, and assistant-driven workflows.

In short, GuardiaPass is an AI-powered security command center: protect, verify, assist, and educate from one place.

Gratitude is a neighborhood discovery platform designed to make local support and opportunities easier to find, trust, and act on.

The app organizes content into tabs for events, volunteer opportunities, food banks/donation resources, community organizations, help & support services, nearby connections, and civics/politics.

Experience is map-first and filter-driven (distance/relevance/date) with list/map/split views, source attribution, confidence indicators, and deduplication safeguards.

Platform remains useful without AI via scraper/API pipelines and rule-based categorization; AI is an enhancement layer for structuring/classification/gap filling.

Includes built-in AI assistance and language customization.

Developed a proof-of-concept background keylogger for cybersecurity research and demonstration.

The system can be deployed via USB, runs in the background, captures keystrokes, stores logs locally, and transmits encrypted reports via scheduled email.

Purpose: demonstrate endpoint security risk, persistence techniques, and exfiltration behavior to support detection and mitigation understanding.

Led complete deployment/restoration of CTIS computer lab infrastructure from non-functional state to fully operational enterprise-style environment.

Diagnosed/repaired/reconfigured Dell PowerEdge and HP servers; built Proxmox cluster for virtualized domain services.

Designed/deployed AD DS, DNS, DHCP, Group Policy, with primary and backup domain controllers.

Deployed ~12 domain-joined workstations with reimaging, drivers, software, and domain integration.

Built accounts/security groups/permissions for controlled access and system security.

Installed rack infrastructure, power/network organization, Ethernet cabling, and tested network stability.

Worked with IT to deploy/configure wireless AP and secure domain-managed connectivity.

Performed extensive hardware diagnostics and restored previously unusable systems.

Strong teamwork with partner enabled successful lab build and deployment.

Team designed QuickVest for Fidelity with focus on accessibility, trust, and usability, reducing onboarding friction for new investors.

Color/contrast choices prioritized visual comfort and inclusion for color-vision accessibility.

Main technical challenge: Figma lacks native external AI integration (Gemini) and existing widgets were insufficient.

Built a custom widget from scratch for Gemini integration in Figma (Edit Mode only; Figma limitation in Presenter View).

Widget is the only black-box AI part; all other flows/pages/designs were built by the team.

Contributed to autonomous environmental cleanup robot for shoreline waste monitoring/removal.

Work included hardware integration, system setup, sensor testing, AI training, and autonomy control logic development.

Used PyTorch, TensorFlow, and Hugging Face resources. Deployed system on NVIDIA Jetson Nano for sensing/processing/autonomous runtime.

Developed Java + Maven steganography analysis tool supporting images/audio/documents (PNG, BMP, WAV, PDF, DOCX, TXT).

Implements LSB techniques for images/audio while preserving file integrity.

Document detection includes PDF metadata analysis, DOCX hidden text formatting checks, and TXT zero-width character detection.

Integrated Apache PDFBox, Apache POI, Java ImageIO, and Java Sound API.

Developed cybersecurity/digital forensics tool that detects fake/renamed files by binary signatures rather than extensions.

Verifies integrity with MD5/SHA-1/SHA-256 and extracts metadata/structure/system properties.

Includes hex-level viewing/search, entropy and MIME analysis, structural validation, and anomaly detection.

Also performs image steganography analysis and hidden data extraction for forensic reporting.

Runs on macOS and Windows; fully implemented in Java.

Built distributed clusters using Raspberry Pi and Turing Pi 2 board with Raspberry Pi CM4 modules and NVIDIA Jetson Nano.

Deployed Docker Swarm and K3s Kubernetes clusters to compare orchestration performance, scalability, and resource use.

Tested master/worker reliability and workload distribution.

Added Ansible automation, Stress-ng load testing, and Prometheus/Node Exporter monitoring.

Configured USB WiFi adapter with monitor mode and packet injection capabilities for wireless security analysis.

Used system for signal broadcasting/monitoring, packet capture, and nearby WiFi analysis.

Demonstrated rogue AP and wireless attack patterns in controlled environments and deepened defensive strategy understanding.

Time Value Vault provides loan payment estimates based on interest, years, payments/year, and initial loan value.

Returns per-payment amount and total interest across loan lifecycle and sends results via email.

Includes finance chat where user asks finance questions and Gemini AI answers.

Designed and implemented a smart doorbell using Raspberry Pi 4 with camera, USB microphone, 3.5mm speaker, button input, and 3.5-inch LCD status display.

Developed in Python (VS Code), based on open-source code adapted for Python 3 and Pi 4/5, configured for startup automation.

GPIO button triggers capture pipeline: libcamera video + ALSA audio; devices configured/tested with Linux CLI.

Integrated Jitsi Meet for real-time audio/video communication and EmailJS notifications with session link on button press.

Work included Linux configuration, peripheral driver setup, camera/ALSA enablement, recording/playback validation, and robust semi-autonomous startup.

(Very similar to a Ring camera workflow.)