Human-centered AI and spatial computing researcher-engineer
I build and study intelligent systems for people learning physical, high-stakes skills, especially when an AI has incomplete evidence about what is happening.
I am completing a B.S. in Statistics & Computer Science at the University of Illinois Urbana-Champaign (expected December 2026; UIUC GPA: 4.00/4.00). At Carle Illinois College of Medicine, I develop mixed-reality and AI systems for procedural medical training. My work spans research prototypes and human-subject evaluation through production-minded Swift, Python, and web systems.
Portfolio | LinkedIn | Email | All merged pull requests
My current questions sit at the intersection of human-computer interaction, AI, and extended reality:
- How should an AI tutor intervene when visual and spatial evidence is incomplete or uncertain?
- How can multimodal sensing support trustworthy feedback for physical skill learning?
- How should a shared AI mentor support multiple learners without reducing agency or participation?
I explore these questions through mixed-reality medical simulation, multimodal system building, and empirical evaluation with learners.
-
Should the AI Speak First? Evaluating Proactive vs. Reactive Facilitation in Mixed-Reality Medical Training
CHI 2026. A 22-participant study of how intervention timing, task phase, and learner preference shape AI facilitation in mixed-reality medical training. -
SpatialTutor: Object-Aware Mixed Reality Training for Procedural Medical Skills Training with AI-Driven Support
IEEE AIxVR 2026. An object-aware training system that connects spatial guidance and AI support to a physical simulation manikin. Best Paper Honorable Mention.
| Project | What it demonstrates | Core technologies |
|---|---|---|
| Lumbera | Apple Swift Student Challenge Winner (2026). An independently designed iPad environment for anatomy learning, procedural guidance, and spatial rehearsal before supervised lumbar-puncture practice. | Swift, SwiftUI, SceneKit, AI agents |
| ImpactLint / Live demo | Reviews warehouse schema changes against DataHub metadata, traces downstream impact, and generates evidence-backed migration artifacts. A measured live run reduced guarded reasoning context by 67.6%. | Python, FastAPI, React, DataHub MCP, SQLGlot |
| StudyCast | A macOS capture desk that runs multiple AirPlay receiver stations for research sessions, teaching labs, and demos, with local previews and per-station recordings. | SwiftUI, GStreamer, FFmpeg, UxPlay |
| Jerry's Room / Live site | An interactive Three.js room whose monitor runs a working React desktop, combining real-time 3D graphics with a usable portfolio interface. | TypeScript, React, Three.js, Framer Motion |
As of August 2026, I have 15 merged pull requests in public open-source projects outside coursework. Selected work includes:
- TestSprite CLI: 12 merged fixes across authentication, timeout handling, validation order, dry-run safety, pagination, and CLI guidance.
- Obsidian Second Brain #175 and #176: a reliable reindex command and Simplified Chinese trigger phrases.
- Collective AI Tools #281: preserved compact-card navigation affordances on touch devices.
These contributions reflect the way I like to work: understand the surrounding system, add focused regression coverage, respond carefully to review, and leave shared code easier to trust.
- Spatial computing: Swift, SwiftUI, RealityKit, ARKit, visionOS, SceneKit, Three.js
- AI and data: Python, FastAPI, multimodal AI, LLM agents, SQL, R, statistical modeling
- Web and systems: TypeScript, React, C++, Java, WebSocket, MCP, cloud platforms
- Research: mixed-methods studies, behavioral analysis, interviews, usability evaluation
I am interested in graduate research and software or research engineering opportunities in human-centered AI, spatial computing, multimodal systems, intelligent learning tools, and reliable AI assistance.