AI Engineering
Building practical ML and LLM-powered applications with an emphasis on reliability and user impact.
Mason Baloun
1st Place Winner – Google AI Startup Hackathon
Building generative AI systems, XR applications, and edge computing solutions. AI security advocate on Capitol Hill.
I'm a Computer Science student at UW–Madison working on vertically integrated AI systems—generative models, spatial computing, robotics, real-time computer vision. My work has been recognized at national competitions and policy forums. I collaborate with researchers from MIT and OpenAI on AI security and governance.
Resume
Core Focus
AR/VR Development
Designing immersive learning and interaction experiences using Unity, XR tooling, and real-time feedback systems.
Robotic Integration
Building end-to-end robotics systems combining edge AI, real-time control, and hardware integration across embedded platforms.
Featured Projects
Beyond Words
Beyond Words is a mobile-first, gamified language learning platform powered by generative AI. Built in 24 hours for a Google AI Startup Hackathon, it won 1st Place among competing teams. The platform teaches German by transforming vocabulary into collectible strategic cards. Instead of static lessons, generative AI dynamically generates contextual prompts, evaluates grammar and sentence structure, scores strategic word usage, produces pronunciation audio, and adapts difficulty in real time. Validated by Google engineers as a production-ready AI content pipeline and adaptive education system.
Next.js React Gemini Firebase Generative AI
Rover XR
Rover XR is an immersive extended reality interface for controlling a physical rover while running real-time object detection on edge hardware. The system sends real-time UDP control packets from XR headset, runs YOLOv8 object detection on a Rubik Pi, streams annotated video back to the headset, and communicates movement commands via UART to rover MCU. A vertically integrated XR + hardware + edge AI system recognized with Honorable Mention at MadData26.
Unity YOLOv8 XR Edge AI Robotics
MotionVault
MotionVault is a mobile-first motion capture platform for robotics and AI training. It converts smartphone camera input into skeletal landmark data using on-device pose detection. Features real-time skeletal overlay, camera-to-storage motion pipeline, vault UI for organizing recordings, export and sharing functionality, and future cloud integration. Transforms consumer smartphones into portable motion capture systems for robotic and AI applications.
Kotlin ML Kit Mobile Robotics AI
PEAR
PEAR is an AR piano learning system built for Meta Quest 3. It overlays spatially anchored note guidance onto a physical keyboard, replacing static sheet music with immersive embodied interaction. Demonstrates spatial computing for embodied skill acquisition and immersive music education.
Unity Meta Quest 3 XR Spatial Computing EdTechRecognition
Presented on Capitol Hill – AI Security & Policy
Board Member – Wisconsin AI Safety Initiative (WAISI)
Presenter – Center for AI Security Advanced AI Expo
Competed nationally against teams from universities including MIT and Harvard. Advanced to finalist rounds and earned a fully funded trip to Washington, DC.
Presentation Topic: AI Agents of Chaos: AI-Enabled Fraudulent Calls
Demonstrated how multi-agent language model systems create new attack surfaces, voice models can mimic emotional speech, AI tools can flood 911 dispatch systems with fake calls, and coordinated AI-generated calls could mislead congressional offices. Built a live AI Hotline simulation to demonstrate emergency system disruption risks. Met with Congressional staffers, engaged Rep. Bill Foster (D-IL-11), collaborated with MIT AI Alignment members, and met one-on-one with Scott Aaronson, OpenAI researcher.
Featured by UW–Madison School of Business
Innovator Showcase – Tech Exploration Lab
Executive Presenter – UW School of Business Event
Featured in UW article while demonstrating PEAR at Tech Exploration Lab event. Presented to cross-industry executives from Google, Salesforce, and Nike.
Presentation Focus: Spatial Computing for Embodied Learning – PEAR AR Piano System
Notable mention: Mike Grall of Fulcrum Solutions observed PEAR demonstration and engaged in technical discussion about the system's architecture.
Let's Connect!
