Unlock full details

Use your master password or a shared access code to view the full projects and resume.

or

Access expires after 24 hours. You can lock it again at any time.

Projects

My projects

Personal projects, team projects, and entries from WSU CougHacks.

These public summaries use generalized names and leave out identifying details. Email me for full project details and my complete resume, or if you’ve been given an access code.

9 projects

Applications Team project

Campus noise map and monitoring

Awarded First Place at the WSU CougHacks hackathon · 2026

A campus app that uses microphones to show noise levels on a floorplan, helping students find quiet or active areas.

  • Web applications
  • IoT
  • Audio & video
More details

Connected IoT devices supply audio and video streams. Security staff have a separate view for monitoring designated campus areas.

Applications Team project

Messaging for care-home residents

Awarded First Place at the WSU CougHacks hackathon · 2025

A web and mobile messaging app for care-home residents to talk with friends and family.

  • Web & mobile
  • Messaging
  • Privacy
More details

Filters personal information out of messages before saving them to the database, so those details aren’t passed to the recipient.

Applications Team project

Habit tracker

Awarded First Place at the WSU CougHacks hackathon · 2023

A team-built web app for tracking habits, with a SvelteKit interface and Flask API.

  • SvelteKit
  • Flask
  • PostgreSQL
More details

Stores habit data in PostgreSQL and was deployed using cloud services.

AI Working prototype

AI language practice

A prototype for practicing spoken Japanese with an AI character that responds out loud. It runs locally and also has an English mode.

  • Python
  • TypeScript
  • Three.js
  • Local inference
More details

Uses speech recognition, a local language model, and speech synthesis to animate a 3D character in the browser. Pronunciation scoring and more languages are planned.

AI Personal project

Text-to-speech for voice chats

A self-hosted tool that reads typed messages aloud in a voice chat. Participants can provide an audio sample to have it speak in their own voice.

  • TypeScript
  • Whisper
  • Qwen3-TTS
  • SQLite
More details

Runs transcription and speech generation locally and queues playback for each channel. Participants can remove their saved voices. GPU memory is freed between speech-generation jobs.

Systems In development

Linux IP failover service

A Linux service for moving a shared IP address between machines. It checks whether peers are healthy and elects which machine should hold the address.

  • C++23
  • Linux
  • Networking
  • Next.js
More details

Includes a web dashboard, dry-run mode, and options for quorum and self-fencing. Still in development; failover needs testing on the intended hardware.

Applications Team project

Home inventory app

A team project for tracking household items, their photos, and where they’re stored.

  • Next.js
  • Flask
  • MongoDB
  • Python
More details

Uses Next.js for the web interface, Flask for the API, and MongoDB to organize items and containers. Photos are stored in a separate image service.

AI Experimental

AI voice chat bridge

An experiment connecting a group voice call to a local speech-to-speech model so it can listen and respond while people are talking.

  • TypeScript
  • WebSocket
  • FFmpeg
  • GPU inference
More details

Mixes call audio, converts sample rates, and sends it over WebSocket. The model server runs, but latency, echo, and turn-taking still need testing on hardware.

Systems Hardware research

Network transceiver diagnostics

Python tools for reading, backing up, and comparing a network transceiver’s EEPROM data with a USB programmer.

  • Python
  • USB
  • EEPROM
  • Diagnostics
More details

I wrote and verified a compatibility change, but the switch still wouldn’t establish a link. I documented the result and how to restore the original data.

The homelab

My Proxmox server

I use a rack server to host local AI models, media services, and ZFS storage. It also runs my home network and VMs for learning Kubernetes.

View my homelab
Server processors
2 × Intel Xeon E5-2690 v3
Compute
24 cores / 48 threads
Memory
128 GB DDR4 ECC registered · 2,133 MT/s