Caden Greenhalgh

CadenGreenhalgh

Computer Science student building software, machine learning, and embedded systems, from applications down to the hardware.

Status

Available for summer 2027 internships

Based in

Anchorage, AK

About

I'm a Computer Science student at the University of Alaska Anchorage with interests in software development, machine learning, computer vision, computer hardware, and embedded systems.

Outside of school, I work as a Genius Admin at Apple, where I troubleshoot and repair hardware and software, solve technical problems, and help people better understand the technology they use.

Beyond technology, I'm into fitness, personal finance, and philosophy. I'm always looking for ways to improve how I learn, think, and work. I like challenging myself with new projects, exploring unfamiliar topics, and building things that are both practical and technically interesting.

School
University of Alaska Anchorage
Degree
B.S. Computer Science, 2027
Interests
ML, computer vision, embedded, networks & distributed systems
Based in
Anchorage, AK

Resume

3 entries

  1. 2026 to Present

    Genius Admin

    AppleWork

    • Orchestrate end-to-end repair workflows and schedule queues for a team of 15 technicians.
    • Supervise daily operational activities within the Repair Room
    • Collaborate cross-functionally with the technical and front-of-house teams to ensure adherence to repair turn-around-time service level agreements.

  2. 2024 to 2026

    Genius

    AppleWork

    • Troubleshot and repaired hardware and software across the Apple ecosystem.
    • Utilized proprietary tools to help solve complex technical problems.
    • Educated individuals on how to use their devices and software more effectively.

  3. 2023 to 2027

    B.S. Computer Science

    University of Alaska AnchorageEducation

Please email me for a copy of my resume.

Projects

4 in progress

Currently building

  1. Roots

    Plant monitoring, from sensor to phone

    C++, Python, FastAPI, TypeScript, React Native

    • C++ sensor firmware publishes readings over MQTT.
    • A FastAPI service stores readings in PostgreSQL and runs in Docker.
    • A React Native app displays the readings on your phone.

  2. gDash

    A personal dashboard for macOS and iOS

    Swift

    • Brings school, personal finances, health, and habit tracking into a single app.
    • Includes a built-in journal for daily entries.

  3. 1Take

    A video-journaling social app

    JavaScript, TypeScript, React Native, Expo

    • Users record each entry in a single take and file it into a journal, limited to one entry per day.
    • Journals can be private or public, and users can subscribe to public journals.

  4. Visual Assistant

    A wearable visual assistant for blind and low-vision users

    Python, OpenCV

    • Runs YOLO object detection on a live camera feed to recognize objects around the user in real time.
    • Converts detections into each object's relative position (left, center, right) and a structured description of the scene.
    • Modular pipeline built to add OCR, depth estimation, and audio feedback.

Previously built

  1. gNews

    A world map of each country's daily headlines

    TypeScript, Next.js, React, D3

    • Zoomable SVG world map: click a country to see its top 10 English-language headlines, with 90 days of history.
    • Pulls headlines from Google News RSS, falling back to the GDELT API for past days, with CDN caching on API responses.
    • Ask AI chat extracts an article's text with Readability and streams a Gemini answer grounded in it.

  2. Dragonfly

    A distributed repair queue for Apple devices

    Python, RabbitMQ

    • Technician stations on any machine on the network create, take, and close appointments through a shared RabbitMQ broker.
    • Routes tickets to separate Mac and mobile priority queues; express appointments go first, and Mac stations pull mobile work when their queue is empty.
    • A ledger service records every event from a fanout exchange and answers station queries over RPC.

  3. NBA Game Outcome Predictor

    End-to-end ML pipeline over 31K+ historical games

    Python, Pandas, scikit-learn, XGBoost, TensorFlow, Keras, Optuna

    • Engineered rolling performance, Elo rating, rest-day, and back-to-back features while preventing temporal data leakage.
    • Built and evaluated Logistic Regression, XGBoost, ARIMA, CNN, and LSTM models with chronological splits and time-series cross-validation, raising accuracy from a 55.3% baseline to about 65%.
    • Tuned XGBoost with Optuna and built an attention-based LSTM over 10-game team histories.
    • Combined models with out-of-fold stacking, keeping training and test data strictly separate.

  4. Chess+ MCTS Opponent

    An AI opponent for a custom chess variant

    GDScript, Godot

    • Built an AI opponent for Chess+, a custom chess variant, using Monte Carlo tree search.

  5. Mini Compiler

    A compiler and interpreter for a small language

    C++, Flex, Bison

    • Lexer and parser built with Flex and Bison, plus an interpreter that executes the parsed program.
    • Supports arithmetic, variables, assignment, trigonometric functions, print, and if, while, for, switch, and break.
    • Runs programs interactively or from a file.

Stack

Tools I work with

  • Daily use
  • Proficient
  • Familiar

Languages

  • Python
  • TypeScript
  • C++
  • Swift
  • JavaScript
  • SQL

Apps & web

  • React Native
  • Expo
  • React
  • FastAPI
  • PostgreSQL

ML & vision

  • scikit-learn
  • XGBoost
  • TensorFlow / Keras
  • Pandas
  • OpenCV
  • YOLO

Systems & tools

  • MQTT
  • Docker
  • Flex / Bison
  • Git
  • Linux

Blog

Coming soon

Currently writing and researching interesting things... come back soon!

Contact

Have a role, a project, or a question? Write to me.

hi@greenhalgh.work

Or send a message here

Answer three prompts and it goes straight to my inbox.

$ contact --to caden

Press Enter to continue